This transcript was generated automatically and lightly edited for readability.

Oliver Strimpel: This is Geology Bites with Oliver Strimpel. After months of high earthquake activity, a fissure opened up near the southwestern tip of Iceland, on the Reykjanes Peninsula, on March 19, 2021. Over a period of about seven months, several other fissures opened up, generating lava flows several kilometers long that filled valleys and created a new 150-meter-high mountain, a sort of mini shield volcano. The eruption has been intensively studied by geologists, because it is the first eruption of its kind in Iceland in living memory, and also because it's extremely accessible, just a 25-minute drive from Reykjavik.

Ed Marshall is one of those geologists, having made dozens of visits to the eruption site to sample the lava and analyze it in his lab. He's a postdoctoral research fellow at the University of Iceland. Ed Marshall, welcome to Geology Bites.

Ed Marshall: Hi, Oliver. Thanks for having me on the program.

Oliver Strimpel: In my introduction, I said you'd visited the eruption site to sample the lava. How do you do that without incinerating yourself?

Ed Marshall: Ha ha. We wear protective gear. So we have a leather, full-body, heavy-duty welding outfit, and we have a face shield from a foundry supply company. And so when we wear all of this together, we're quite heat-resistant. And actually, you have to be sure to only spend a short time, because you can get too comfortable near the lava, and parts of your body can get too hot. Typically we don't try to stay more than 10 seconds, or 15 seconds at most, near the lava.

Oliver Strimpel: Wow. With all that gear, and 15 seconds is the max?

Ed Marshall: Very close to the lava, the radiant heat is so intense that burning is an issue. Of course, the gear that we're wearing is completely flame-resistant, but you still don't want to be smoking. There's no reason to spend a long time near the lava. Taking breaks and coming back and sampling the lava again is completely sufficient. The lava is going to be there for several minutes at a time before it hardens. Then the next piece of gear that we have is a long metal pole with a scoop on it, and we stick the long pole into the lava. It has kind of a nice gooey texture. Depending on the lava, it can be extremely viscous, and you'll have to use a lot of force to move it. It's like a very thick tar. And sometimes it's more like molten marshmallow, if it's very gas-rich and very fluid, and so it's very easy to sample. So we'll scoop out the lava, and we'll immediately dunk it into a bucket of water.

Oliver Strimpel: So I suppose it depends a bit on how active the eruption is at the time, where exactly you can get to, and where you go to sample the lava. It's been very variable, hasn't it?

Ed Marshall: Oh, extremely variable. So this eruption started with one small vent producing some nice lobes of lava that were very easy to sample, but progressively vents opened up, and then there was a very high fountaining episode, and the lava field expanded incredibly. And so on some days you would be able to easily sample. Some days are just extremely easy: you pull up with your car, it's a short walk, you scoop the lava, you're in and out in 20 minutes. Some days it's much harder, or impossible. The lava may be coming out of the volcano and not reaching the edge, and at such a young volcano, you don't want to walk across older lava, because the older lava isn't very old here. It's only weeks old. There's still an extreme danger of falling through the lava.

Oliver Strimpel: Why did you make so many visits to sample the lava? Why not stop when you got, say, I don't know, a couple of good-sized samples, a few kilograms?

Ed Marshall: Something that we found very early on is that the lava is changing in composition in time. So why did we make so many visits? We want to see this change in time in the lava composition. We've been very successful at sampling frequently enough that we're really able to map out the composition over time. It's just been wonderful.

Oliver Strimpel: Oh, OK. So you said you had this nice long pole with a scoop at the end, and as soon as you grab some lava, you dunked it in some water. Why did you pour it into water right away?

Ed Marshall: We pour it right into water to quench it into a glass. So as lava cools, it will crystallize, and this is actually not a good thing for geochemists, because if you're interested in the composition of the liquid melt, then if it crystallizes, that liquid melt has now disappeared. It's turned into crystals. And so we scoop the lava and quench it, and the water will cool it so rapidly it will turn into glass and won't crystallize. And this glass is excellent for in situ analysis, and excellent for analysis of the glass directly, particularly for looking at volatiles that would be lost as crystals are formed.

Oliver Strimpel: Oh, I see. So if you allowed it to crystallize, you'd get some partitioning, I guess, of the lava into what's in the crystal and what's still in the remaining melt, and you'd start to see variations in composition depending on which phase you actually wound up measuring. But you want to see what the bulk composition of the lava is. So that's why you make it into a glass, is that right?

Ed Marshall: Yes, that's part of the reason. Many of the types of analyses that we want to do are microscale in situ analyses, where we use a beam of electrons, or secondary ions, that are microns across. And so what we want to do is preserve as much of the glass as possible, so when we measure the sample with a secondary ion beam, we can directly measure this liquid material. So we quench it, and so it preserves the greatest amount of geochemical information about the melt itself.

Oliver Strimpel: OK, I'm going to ask you about the analyses that you've done and what you found out. But before we talk about that, let's talk a little bit about the scientific questions you're tackling.

Ed Marshall: So early on in the eruption, we were mostly interested in the eruption as a phenomenon, particularly from a hazard perspective: what was it going to do next? To do that, you want very basic information very quickly. How much water's in the melt? How much volatiles are in the melt? Is this amount changing over time? What we found is that it is changing, and this may have led to a change in eruptive style over the course of the eruption: from early on, a very effusive eruption, with no fountaining and very little ejected lava at all, and later in the eruption, quite a bit of fountaining of lava. But the eruption never became explosive.

But we're also interested in where the lava is coming from. What's happening beneath the volcano? And what kinds of processes led to this eruption and are happening concurrently during the eruption? We're interested in any changes in the lava. We're interested in the crystal cargoes that can tell us about how quickly the melts come to the surface and how long the melt was situated beneath the volcano. And we're interested in putting this eruption in perspective relative to the other eruptions along the Reykjanes. And then finally, it's quite likely that this eruption marks the beginning of an episode of eruptions within the Reykjanes. Previous studies have shown that the Reykjanes erupts in episodes of volcanic activity, each one about 400 years long, spaced about 800 years apart, or a thousand years apart. The last eruption on the Reykjanes was about 800 years ago, and so in that sense, we're kind of due for the next volcanic episode. So there's a lot of interest in this eruption, because it will probably be an example of the eruptive style that we'll see over the next few hundred years.

Oliver Strimpel: So how can a detailed study of the erupted lavas throw light on these scientific questions?

Ed Marshall: So a lava has different parts, and each one of these parts can tell a different side of the story. Probably the hallmark of an eruption is liquid melt coming out of the ground. The liquid lava can tell you a lot about how the composition of the lava is similar to or different from other lavas in the vicinity. It can tell you about how much volatiles are coming out of the volcano. It can tell you about the types of changes that are occurring beneath the volcano, in the magma chamber that supplies the eruption. And then you can look at the crystals, the crystals that are brought to the surface. They can tell you about the storage history of crystal cargoes and the crystal mush that lies beneath the volcano. They can also tell you about how quickly the lava ascends to the surface during the eruption. We're also interested in the trace element and isotopic data that is trapped in the melt, and this can tell you an enormous amount about any mixing processes that are occurring, and also it can tell you about the types of melt that are supplying the volcano from the mantle source. And this is something that our group actually focuses on.

And then the xenoliths that are brought up in the eruption. So xenoliths are fragments of rock that are carried by the melt to the surface, and these can be portions of the crust, or they can be material from the magma chamber. These rocks preserve a lot of information, not only about the eruption, in the case of fragments of rock from the magma chamber, but they can also tell us about the deep crust of Iceland. One of the really interesting things about the lower crustal xenoliths is how deformed they are. Of course, the Reykjanes is lying along a plate boundary between the European and North American plates, and the xenoliths accommodate an enormous amount of strain that occurs from the sliding of the two plates beside each other. They're being stressed, and they are deformed as the plates move, and we see a lot of deformation in some of the xenoliths. So this is something that we're also interested in: how deformation is accommodated in Iceland.

Oliver Strimpel: Lot of things to look for. Let's pick out one at a time. How do you measure the trace element abundances, and what do your results show?

Ed Marshall: We dissolve the rock in hydrofluoric acid, and we make a solution that is then run on a mass spectrometer. Trace elements are very low-abundance elements, like the rare earth elements, or rubidium, or niobium, or tantalum, that are in the rock. They're in very low abundance, but with mass spectrometry we can measure them.

There are two kinds of processes that give rise to these heterogeneous melts. One of these is lithological and chemical heterogeneities within the mantle. Just like you go out and you look at the crust on a walk, and you look at the rocks, and there are lots of different rocks, well, there are lots of different rocks in the mantle as well. It is not just homogeneous peridotites. But also, as the mantle upwells beneath Iceland, or upwells beneath Hawaii, or any hotspot or rift, it will gradually melt as it ascends, and the melts that are produced first are very different from the melts that are produced at the end of melting. And this is true even if you're melting something that's completely homogeneous. So we have progressive change because of partial melting on one hand, and we have physical, lithological heterogeneities that exist in the mantle on the other hand, and these give rise to a very wide spectrum of melt compositions that come to the surface. And so trace elements are very good at sniffing out the differences in the components.

Oliver Strimpel: You said that there's another observation that bears critically on this, and that's the measurement of isotopes. So let's talk a bit about isotopes. How do you measure the isotope ratios, and what do they tell us?

Ed Marshall: You dissolve the rock like you would do for trace elements, but then you need to isolate the element of interest. So we have a series of chemical procedures that involve chromatography. So you take your rock and you put it through a chromatographic resin. The resin is able to separate different elements from one another. Once we have our element of interest isolated and in a solution (so we'll end up with a solution, for example, of nitric acid with just strontium in it), we can take the solution and we can put it into a multi-collector mass spectrometer or thermal ionization mass spectrometer, which are mass spectrometers that have many collectors and are able to detect not only the signals of a single isotope, but all masses from the particular element simultaneously.

Isotopes are essentially only sensitive to these lithological and chemical heterogeneities that I talked about with trace elements. So the difference is that isotopes are not sensitive at all to this progressive melting that occurs over time. If you have a homogeneous mantle, then you only end up with one isotope composition. But because we have a heterogeneous mantle, we detect many different isotope signatures coming out of a particular eruption. And this is what we see in the Fagradalsfjall eruption. And actually, we see quite a bit more variability than that. We see it swing from one place to another and back. And this is telling us that we have mixing within the magma chamber, and so the source that we're sensing keeps changing as mixing swings us from one type of melt to another type of melt.

Oliver Strimpel: That's fascinating. So the isotopes give you a unique handle on the lithological origin, or the inhomogeneities of the actual origin, and the trace element compositions give you essentially the melt fraction, because you can actually disentangle the effects of partial melting and heterogeneous source material using those two types of measurement.

Ed Marshall: Yeah, theoretically. In practice, of course, it's quite difficult, because in the trace elements, the melting and lithological heterogeneity signals are intertangled in a way that makes them difficult to deconvolve. But yes, in theory, you can disentangle these influences.

Oliver Strimpel: Because when we spoke about it, you said that even if you think of two sources that are heterogeneous, they themselves will be mixtures, perhaps, of other sources. And so you're talking about mixtures of mixtures, and that all produces complications that you need to disentangle with all kinds of fancy statistical methods, I suppose.

Ed Marshall: Exactly. There's a concept in mantle geochemistry of an instantaneous melt. So when you have a tiny parcel of the mantle and you melt it a little bit, you'll produce this tiny amount of melt that we call an infinitesimal melt. But the infinitesimal melts are real, in that the mantle melts, and as it is melted, the melt is extracted. And so in a sense, erupted melts are made out of the integration of all of these infinitesimal melts together. And what I mean by the integration, physically, is mixing. So these melts are really the mixture of a vast array of wide compositions of melts coming out of the mantle. And what we see beneath Fagradalsfjall is now two melts, themselves representing mixtures of a vast array of melts, that are still quite different compositions, that are now mixing with one another. And the neat thing is, now we can see this change as these two bodies of melt mix together.

Oliver Strimpel: I want to just digress a second and ask you to pronounce clearly the name of this volcano, the name of this eruption.

Ed Marshall: Yeah, it's Fagradalsfjall.

Oliver Strimpel: OK. All right. Thank you. So we talked about trace element concentrations. We've talked about isotope ratios. What about the crystals and the xenoliths that you find in the lava? What do they tell us?

Ed Marshall: So crystals and xenoliths tell us an intertwined picture. Some of the crystals are within the melt because they had fallen off of xenoliths, but most of the crystals that we see within the lava have not. The crystals contain melt inclusions, or tiny pockets of melt that were trapped inside the crystals as the crystals grew, and by analyzing those melt inclusions, we can learn about the melts that these crystals grew from. And what we see in these melt inclusions in the crystals is a wide spectrum of melts, a very, very large population of melts, from compositions that are much, much, much more depleted in incompatible elements than the eruption we see today, to much, much, much more enriched in incompatible elements. So what I think is that these melt inclusions are sampling melts that have mixed together to form the melt that is currently erupting beneath Fagradalsfjall. And this wide array of melt inclusions shows us that beneath Fagradalsfjall, at the crust-mantle boundary, it's a dynamic environment, and the melts are moving around and constantly mixing with one another.

Now, the xenoliths are telling a different story, because the xenoliths are mostly coming from the lower crust of Iceland, and these xenoliths have a variety of different textures and petrologies. Some of them are deformed. Some of them have purely igneous textures. Some of them are completely recrystallized and have quite metamorphic textures. The general picture is what you might expect beneath a volcanic plate boundary: it's hot, it's under stress, it's deforming and recrystallizing. And these are the types of rocks that we're seeing in the xenoliths.

Oliver Strimpel: So overall, are you able to tell the depth from which the magma originated? And is it really from the very bottom of the lithosphere, albeit a rather thin lithosphere, Iceland being on a plate boundary, as opposed to a magma chamber within the crust or within the lithosphere, where it was kind of parked, if you like, on the way up?

Ed Marshall: This is a significant finding. It is sourced from a deep magma reservoir that is beneath the crust. And you're right: because this is a rift, the lithospheric mantle is very thin or nonexistent, and so the lithosphere properly begins essentially at the crust-mantle boundary. The crust-mantle boundary represents the first major density boundary, but also the first major boundary in porosity that melts will encounter as they percolate upwards through the mantle. So this crust-mantle boundary is actually a natural place for melts to accumulate at rifts. And so in that sense, it's not necessarily a surprise that this eruption is sourced from the crust-mantle boundary. But it is a surprise in that eruptions from a crust-mantle-boundary magma reservoir are very rare, and this is actually one of the first, certainly the first eruption from a crust-mantle-boundary reservoir that we've been able to monitor very closely. Many rift eruptions, such as at mid-ocean ridges, erupt from the crust-mantle boundary, but of course that's submarine. The Fagradalsfjall eruption gives us a wonderful opportunity to study these mid-ocean-ridge-like eruptions from a dry position.

Oliver Strimpel: Just to be clear, when you say these originated from the crust-mantle boundary, you mean the boundary between the crust and the convective mantle, because there is no lithospheric mantle, or between the crust and the lithospheric mantle?

Ed Marshall: The lithospheric mantle is formed through the cooling and partial melting of the mantle, and because we're at a rift, the lithospheric mantle is probably very thin or nonexistent, simply because it's a point of active upwelling, and there hasn't been enough time for it to cool and form a lithospheric lid.

Oliver Strimpel: Wow, fascinating. So we're actually seeing lava that is coming pretty directly from the convecting mantle below, the asthenosphere?

Ed Marshall: Yes, precisely. So this is an asthenospheric eruption.

Oliver Strimpel: And how deep is it coming from?

Ed Marshall: The thickness of the crust, so 15 kilometers, more than 15 kilometers depth.

Oliver Strimpel: In an earlier podcast, Marie Edmonds talked about volcanic gases and how their composition can reveal the depth of the source magma. For example, the sulfur dioxide to carbon dioxide ratio increases as the magma approaches the surface. This eruption emits gases. How has the analysis of that gas been consistent with your results from your other measurements?

Ed Marshall: Yes. So this has been a point of intensive study during this eruption. Gases are a safety hazard. But like Marie Edmonds said, because this is a deeper eruption, this eruption has an extremely high CO2/SO2 ratio. A high CO2/SO2 ratio indicates a deep source: further support that this volcano erupts essentially directly from the asthenosphere beneath the crust-mantle boundary.

Oliver Strimpel: In another previous podcast, Steve Sparks talked about a giant area of crystal mush at the base of the lithosphere, containing the partial melt, which becomes unstable, eventually amalgamates, and then feeds volcanic eruptions. Is your work on this eruption able to test the validity of that picture?

Ed Marshall: One of the notable aspects of the Fagradalsfjall eruption is that there's an extremely large number of crystals. One possibility is that the crystal-rich nature of the lava is a direct result of decompaction of mush immediately before the eruption. The crystal-rich nature of the lava may be evidence that the mush surrounding the magma body has been disturbed, maybe by magma input, and the crystals have been mixed into the melt. And so at the surface, we see very primitive crystals, out of equilibrium with the melt, that retain melt inclusions that are dissimilar from the erupting melt composition. And this would all be expected if the crystals within the lava were from a mush that has an older history than the lava that's actively erupting out of the volcano today.

Oliver Strimpel: And that mush would be in a layer at the base of the crust, between the crust and the asthenosphere?

Ed Marshall: Correct, which will form as a magma body crystallizes and creates a crystal pile. And this crystal pile won't completely compact to zero porosity. It will be a mush; it will be partly melt and partly crystals. And these mushes can be disturbed by processes like magma injection. And this may explain why we have such crystal-rich magma: magma injection shortly before the eruption, entraining some of this mush material into the magma immediately before it erupted. So you could imagine that there could be a layer of mush that lines the top of the magma chamber, and the conduit to the surface may have punched through this mush, and as the eruption continues, it continuously erodes mush away. And this is another possibility for why this eruption is so crystal-rich, but again, invoking these mush bodies surrounding the magma chamber.

Oliver Strimpel: Do the answers to these questions that we've been addressing, such as the depth at which the lava originates and how homogeneous the source is, and so on, give us clues as to larger questions about the structure of the lithosphere and the nature of the boundary between, in this case, the lithosphere and the convecting mantle below, i.e., the asthenosphere?

Ed Marshall: Well, I think that questions related to the physical structure are probably more directly and more easily answered through seismic analysis. But I think that the geochemical data is telling us the types of processes happening at the crust-mantle boundary. We knew that the crust-mantle boundary should be a place where melts accumulate, but now we can actually see this happening. Now we actually know the timescales on which this happens. We know what kinds of melts are there. Now we know that we do have this extreme spectrum of melts present at the crust-mantle boundary, because they're erupting out of a volcano. We know that these mixing events are very rapid, because we've seen, over the course of a few months, a magma body undergo this enormous mixing event and change radically in its chemical composition. It's amazing to observe very clearly these signals that are coming from a part of the Earth that's typically very difficult to study, and typically studied only indirectly. So you would typically study the crust-mantle boundary via crystals that are brought up in melts, and now we can study them through the melts themselves, and that involves far fewer assumptions than looking at melt inclusions.

Oliver Strimpel: So here we are. We're doing the interview at the beginning of November 2021. The eruption started in March. Is the eruption still going on?

Ed Marshall: There is no lava that is flowing out of the volcano right now, but there is degassing that's occurring. The number of earthquakes is decreasing over time. So currently it's looking like the volcano is shutting down, but there's no reason that it can't start back up again, and there's no reason that another eruption can't happen in the Reykjanes. This is likely the first of an episode of eruptions that will take place in the Reykjanes, if the historical pattern continues. And if this is true, then we should have more eruptions to look forward to, but maybe in 30 or 40 or 50 years or more.

Oliver Strimpel: So if this volcano has indeed shut down for now, how long will it take you to complete the analysis of all the lava that you've already collected?

Ed Marshall: In geochemistry, there's always more geochemical analysis to do. Just when you feel like you've collected enough data, there's another geochemical tool that you want to test out on the rocks that you have. The Fagradalsfjall eruption will be a collection of rocks that many people will spend careers on. The eruption is wonderfully multifaceted. You have this amazing, striking change in melt composition, but you also have an extremely high crystal content of the melt, which has its own story. And then we also have xenoliths that are coming up, which tell their own separate story. I think there's a lot to learn about crust-mantle magmatic systems that exist at the crust-mantle boundary. But how long will it take me? If I were doing the work, it would take me more than my lifetime to get through everything. But the simple things that we talked about on this program, the major, trace and some isotope measurements, will probably be done in the next few months.

Oliver Strimpel: Ed Marshall, thank you very much.

Ed Marshall: Thanks so much, Oliver, for having me.

Oliver Strimpel: For more about Geology Bites, as well as pictures and illustrations that support this podcast, you can go to geologybites.com.