What pathways for ascending magma? Behind the scenes with Eleonora Rivalta

---

Eleonora Rivalta is one of the leading figures when it comes to modelling dykes. Much of her research has been, and continues to be, dedicated to developing numerical models that can simulate the pathways of magma travelling through the Earth’s crust. Eleonora’s models are based on the underlying physics of rock mechanics and the volcanic system, and are able to accurately predict the pathways of dykes — both in the lab (as described in this recent paper), and in nature (read here about how Eleonora’s model was used to forecast magma trajectories at Campi Flegrei).

Back in May, the VIPS commission were delighted to host a guest lecture from Eleonora Rivalta: What pathways for ascending magma? Implications for geophysics and petrology. In her talk, Eleonora described how magma travels through the surface as a very sharp object which can easily fracture the hardest rock. She explained that although there are advanced models that can predict ascending magma pathways in many cases, there’s still a lot left to be understood… and magma can often exhibit unexpected behaviour. To find out more, check out the live stream of Eleonora’s full talk:

Eleonora agreed to an exclusive behind the scenes interview with the VIPS blog team after her seminar. Keep reading to find out more about Eleonora’s research, upcoming work, and why Eleonora thinks we need more collaboration between people of different fields in the VIPS community…

---

What is your research background, and how did you end up working on models of magma transport?

So my background is in physics. I studied a degree in physics in Bologna, and at some stage in the course I had to choose what branch of physics to study. I decided to go for geophysics, and I had classes that I liked very much… on the mechanics of the crust, and the mechanics of solids and fluids and how they interact. Then for my master’s degree I decided to pursue a more mathematical study of dykes, and over time I made the choice to go from a very mathematical perspective to something a bit more applied.

What do you find most rewarding about your research?

I think it took a while for me to find it really rewarding. At the very beginning when I was working on the mathematical models of dykes, I found it rewarding when real life observations matched the mathematics in some way. I liked it because it struck me as a miracle that the mathematical equations could really roughly predict something real, especially in a predictive way and not just matching results in retrospect.

But I think there were two things that led me to enjoy my research very, very much. One thing was when I started doing laboratory experiments of dyke propagation in gelatine. These experiments allowed me to really see the physical processes of dyking with my own eyes, and to understand so many things about them.

The second thing was to go to the field with geologists, which led me to see these processes in the field. At first it was difficult to go in the field with geologists, because geologists are trained to see dykes and exposures with their eyes, and I didn’t have the training to see what other people were seeing. Moreover, the geochemists were using so many words I did not understand. It took a while to feel comfortable in the field, but, after I gained courage and started asking questions, I found it very, very enlightening and I became more and more curious to see more.

I think the whole volcanology community is very important

How can the tools that you have developed be used to help volcanic observatories in crisis situations?

This is a very good question. So one critical point is that, obviously, the more sophisticated a tool is the more difficult it is to use. If you need programming skills just to run and use a numerical model, then the community that can use it is small. Volcano observatories are more trained for monitoring, and installing and running the monitoring instruments, and they don’t often have the time to learn complex codes and adapt them to their needs. Therefore it is very useful when the complexity of a numerical model is hidden so that it can be used by experts without them having to be confronted with the underlying mathematics. This can be achieved, and it’s a very critical skill that some people have.

There are people who are working on coding up the physics and mathematics of magmatic processes, and then there are people who know the needs of the observatories… they know what they are actually looking for. For example, at the end of my talk I recommended two webpages. The first is the webpage of Mehdi Nikkhoo (https://volcanodeformation.com/). He is a scientist who develops fantastic easy-to-use Matlab functions for modelling different magmatic processes. These functions have been used by the French scientist Francois Beauducel to produce a model for the French volcano observatories that uses GPS data to figure out how magma behaves as it ascends.

The second webpage that I recommended is the github page of my PhD student Tim Davis (https://github.com/Timmmdavis), who developed an amazing 3D dyke propagation model. To use this you need some programming skills, but for it to be used by an observatory you just need somebody in the middle who understands both the programming language and the needs of the observatory. Then you need data to be collected to test the code. So that’s why I think the whole volcanology community is very important, it’s important to have the contribution of many people so that the complex codes are meaningful and can be really used.

What would you say is a common misconception regarding magma ascent?

I think there are many and it’s very important to reject them and make it clear. One idea that I tried to address as clearly as possible in my talk is that people very often assume magma needs some kind of pre-existing channel to propagate.

Magma may use a channel if it’s there, but it doesn’t need any. This has several bad implications. If you assume this then you are less able to predict what magma is going to do, because you underestimate the power of magma. Dykes are very sharp, and there are laws that command their propagation. If you think that they are governed by different laws then you are not going to be able to accurately forecast how dykes will behave. For example, if you believe that magma preferentially takes pre-existing channels, then you absolutely don’t understand monogenetic fields. Why should monogenetic fields exist if magma always ascends through pre-existing channels?

As soon as you consider that actually magma can propagate obliquely, everything changes. Peripheral activity of the volcano can actually be fed from the centre of the caldera, and your conclusions become completely different. This misconception can be seen in most pictures – if you google “magma ascent” you will find a forest of many diagrams with a big magma chamber and paths that all go vertically. Again, this assumption means that you are unable to understand your system – if you think the source of magma is located below an eruption, then you are only going to look below. If you know that the source could also be at the side, then you have a better chance of finding it.

What are the biggest challenges in your field?

In the talk I explained that there are basically two types of models for magma ascent. In one model you completely neglect magma viscosity and the fluid mechanics of the magma, and by making this approximation you’re better able to calculate the direction of the magma pathways. The second type of model takes the fluid dynamics of the magma into account, and can therefore predict the velocity of the dyke and the actual dynamics, but only on pre-set pathways.

These two approaches are separate so far, and the challenge is to bring them together so that you can simulate everything at the same time. This would bring the benefit of being able to forecast the magma trajectory and time to eruption, as well as how changes in the magma, such as exsolving bubbles and crystallisation, affect the propagation. We cannot do this yet as we have not brought together the two types of models very well. Along with Francesco Maccaferri, Virginie Pinel and her students and postdocs, and my students Lorenzo Mantiloni and Valentina Armeni, I am working on the MagmaPropagator project to address this problem.

Another challenge is that we need more data to validate the codes, and more advanced techniques to get the data. The predictions from models cannot always be proven. For example, I said earlier that people are used to looking just below an eruption for the source of the melt. But if it doesn’t come from below, if it comes from the side, then some of the tools that we have developed to look for it are not appropriate. There needs to be parallel progress in the data-collection techniques, inspired by the models, as well as progress in the modelling. I think this is another example of why we need integration of people with different backgrounds – the modelers, the field people, the geophysicists – and more and more talking between these people.

---

Do you have any advice for Early Career Researchers who would like to use numerical models in volcanology?

I think maybe the best advice would be to read textbooks and papers. Read the papers that you like, that are written in a style that you enjoy. There are some papers that really make it clear what the open questions are, and whether there is work needed. As you read all this stuff, you may find that there is a gap that interests you and you may think “oh this experienced person says this hasn’t been proven yet… maybe with my code, maybe with my abilities, I can actually do this”, and then you can go and do it.

A lot is about how you read, not just how much or which papers. At first you can skim through and see if it meets your interest, and then read it in a way that you are then able to explain it to somebody else. Try to summarise it, and also think about how you would use it. Would you use this paper in the introduction to motivate your work, or would you use it to support your findings in the end? You can even place the citations in your work as you go along. If you think a paper is good for your introduction, place the citation there as soon as you’ve read it. Then in the end when you have a lot of citations you rearrange them in a logical way, and there you have it… sexy work.

Now that new instruments and models are coming, gravimetry has the potential to estimate the mass flow rate of ascending fluid

Can you give us a glimpse into the next steps in your research? Are there any upcoming papers that our VIPS followers should be keeping an eye out for?

The paper that I presented on the 3D numerical model of propagation has just been accepted, so this will be out soon as Davis et al in Geophysical Research Letters.

Then there is work that my postdoc Mehdi Nikkhoo is doing about gravity. Here I need some context: gravimetry is the only technique that can identify subsurface mass changes. When the distribution of mass changes so does the gravity, and you can detect these changes with gravimeters. Recently there has been huge progress in the instrumentation, and people have found a way to develop miniaturised accelerometers – that are in mobile phones – into cheap gravimeters.

Mehdi has been working on very fundamental theory linking tri-axial dilation, that is: not spherical, not ellipsoidal, but amorphous expansion in rocks, to the ensuing ground displacements and gravity changes. On one hand, this will help with modelling magma chambers, which we know are not spherical nor ellipsoidal. On the other hand, this will make gravimeters much more usable, which really opens the possibility of using gravimetry to identify mass beneath volcanoes.

This may seem very technical, but actually it’s very, very, very important, because if you can identify mass, then you know what fluid you’re dealing with. You can assess the density and then you can tell, for example, if it is a melt with bubbles, a melt without bubbles, a hydrothermal fluid… you really know what it is. This could be very big progress because so far gravimetry has not been used so much for volcanoes due to the instrumental cost, which now has a solution, and due to the lack of theoretical models to understand the observations – which can be very counterintuitive. Now that new instruments and models are coming, gravimetry has the potential to estimate the mass flow rate of ascending fluid, and in theory you could say “Okay, there is a lava fountain coming”. This work is with the Newton-g project – check out their cool video.

What is your favourite volcano?

If you had asked me at the very beginning, I would have definitely said one in Iceland because I am studying dykes, and these volcanoes were simply the most comprehensible for me. Then over time I also dealt with other types… I dealt with calderas, I dealt with dome-forming volcanoes… it’s hard to answer this one, they are all amazing. But I can say Etna and Campi Flegrei… because I’m Italian… so I can go for these two. Etna for the level of activity, which is super varied and always different, and Campi Flegrei because we don’t actually know what’s going on there. They are both a huge challenge in different ways.

---

Eleonora is currently based at the GFZ German Research Centre for Geosciences, Helmholtz Centre Potsdam, where she is the leader of the Magma Propagation research group. She is also an associate professor at the University of Bologna in Italy. Since the time of interview, Eleonora’s recent paper on 3D magma propagation led by her PhD student Timothy Davis has been published. Click here to check it out!

---

---

Would you like to be featured on our blog?

We’re on the hunt for fellow VIPS enthusiasts to share their stories, whether you’re a student, an early career researcher (ECR) or more experienced researcher – we want to hear from you!

Do you have a new paper to tell the world about? Or have you been involved in some exciting fieldwork recently? Get in touch at info@vipscommission.org to tell us what’s going on in your field of VIPS.