February 20, 2013

LASI V: Silica Flour Crust and Vegetable Oil Magma– Using Laboratory Experiments to Better Understand Ground Deformation, Magma Movement, and Volcanic Eruptions

Posted by Evelyn Mervine

Note: Dr. Olivier Galland, a senior researcher at the Physics of Geological Processes Center of Excellence at the University of Oslo, presented a talk, “Ground deformation associated with shallow magma intrusions” at the LASI V conference in Port Elizabeth, South Africa in October 2012. The article below is based on this talk and also an interview with Dr. Galland. Over a few weeks, I am highlighting some of the research that was presented at the LASI V workshop. This is the fourth post in that series.

The landscape around an active volcano is dynamic: as magma moves about, the ground may change shape, inflating or deflating depending on how the magma is moving (Steps 1-2 of Fig. 1). The deformation may be caused either by the deep movement of magma or by the shallow intrusion of magma as sills or dykes. The movement can be dramatic (hundreds of meters) or it can be barely noticeable (a few millimeters). Big or small, ground deformation at volcanoes is important to monitor as it can provide clues to what is going on under the surface and, importantly, about when and where a volcanic eruption might occur.

At many volcanoes, geologists monitor ground deformation using a variety of techniques, including physical devices such as GPS and tiltometers, remote sensing techniques such as InSAR, and photogrammetry, which is the analysis of photographic images (Step 3 of Fig. 1). From surface measurements of ground movements, geologists can try to model what is going on underneath the surface (Step 4 of Fig. 1). However, linking observations of surface deformation to magma movements is no easy task (Step 5 of Fig. 1). That’s because geologists cannot generally directly observe the insides of a volcano. An analogy might be trying to understand why a kitchen sink is broken—why it is leaking or not flowing, perhaps— without being able to open up the cabinet door and look at the sink’s plumbing. Geologists can observe ancient volcanic plumbing, so to speak, in places where the tops of volcanoes have eroded away. However, observing the plumbing of active volcanoes is very difficult to impossible. Therefore, geologists generally rely on fancy computer models to try to link surface deformation with subsurface magma movement.

Fig. 1. Schematic diagram illustrating the principle of ground deformation analyses on active volcanoes. Numbering gives the succession of the stages of the analyses. 1. Magma intrudes in volcano, feeding magma reservoir or forming a sheet intrusion (dark gray intrusion). 2. Magma intrusion triggers ground deformation at surface, leading to modified topography (dashed line). 3. Topography variation is measured by geodetic techniques (GPS, InSAR, Photogrammetry, etc.). 4. Geodetic data are compared with modelling of ground deformation due to various intrusion shapes. 5. The best fit between models and ground deformation data provides a calculated intrusion shape (light gray dashed intrusion) responsible for the measured ground deformation. Nevertheless, the calculated intrusion shape is not a unique solution, and no direct observation is available to validate the calculation results.

Computer models are extremely helpful in trying to understand the insides of active volcanoes. However, they do have their limitations. This is because the shapes of magma chambers and shallow magma intrusions can be complex. Computer models can easily accommodate simple shapes such as circles or flat bodies. However, they have a harder time simulating more complex shapes, such as cones or saucers. Computer models also have trouble taking into account all environmental factors, such as the various stress fields and the realistic behavior of rocks. Finally, it is difficult to be completely confident in the results of computer models since the results of such models cannot generally be compared with direct observations.

One way to try to bridge the gap between surface observations of ground deformation and subsurface movements of magma inferred through computer models is through laboratory experiments. Dr. Olivier Galland is one scientist who conducts such experiments. Dr. Galland describes his experimental set-up (Fig. 2a) as, “Actually, it’s very simple. It’s just a box filled with silica flour, which is a fine-grained, granular material that is simulating the brittle crust. And into that material I inject a vegetable oil, which represents the magma, and the pressure of that oil is monitored. And then it’s possible to simulate the transport of magma and the formation of dykes and sills in this box.” Although such procedure seems simplistic, robust dimensional analysis of the method shows that the experimental results at the lab scale are valid at the geological scale.

Fig. 2. a. Drawing of the experimental apparatus. b. Representative oblique view photograph of the model surface during an experiment. The surface exhibited a smooth relief, at the rim of which the oil erupted. Dashed white line locates the section in the next image. c. Representative oblique view photograph of the model after the end of the experiment. The oil solidified and the intrusion was excavated, such that its top surface can be observed. The section showed that the intrusion was a thin sheet, resulting from hydraulic fracturing. Thus, the top surface of the intrusion was representative of its whole shape.

In a recent experiment (Galland, 2012) Dr. Galland used this silica flour and vegetable oil set up to simulate three types of shallow magmatic intrusions: a cone sheet, a dyke connected to a cone sheet, and a saucer-shaped sill. The motivation for this experiment was to provide additional data about the deformation that occurs when these shallow magmatic bodies are emplaced and, in addition, the deformation that occurs when they breach the surface and erupt. While the silica flour and vegetable oil set-up can’t simulate everything about a real volcano, it can provide plenty of important information, especially about the shapes of various intrusions and the pressures of the magma which produces these intrusions. This is possible because the oil is solid at room temperature, such that the final intrusion can be excavated out of the silica flour and its 3D shape analyzed (Fig. 2c).

Dr. Galland explains, “When a volcano is about to erupt or as it is erupted it either inflates or deflates, and this movement can be monitored. And then afterwards this movement is analyzed and inverted by geophysicists who try to calculate the geometry of the magma conduits. The problem is that we can never test these modeling results because we don’t see the intrusions because they are buried. The only way to really test these inversion techniques is to have a system where you measure the deformation and also know the shape of the intrusion and the pressure into the intrusion. And this is what we are doing in our experiments. We can quantify the shape of the intrusion and measure the pressure into the intrusion and measure the surface deformation, and then we can use this dataset to help test the inversion tools used by the geophysicists.”

Already, Dr. Galland’s experimental work is providing some interesting clues as to what is going on under the surface at volcanoes. “Recently, we figured out that there is clearly a link between how magma rises towards the surface and the pattern of the deformation at the surface,” Dr. Galland says. “Interestingly, this pattern develops very early during the experiments in the very early stages. So, we can observe some asymmetric development of this pattern which indicates where the magma is rising. So, if we can analyze that in real time, we could theoretically use it to predict where an eruption will occur. Obviously, in terms of volcanic hazard that is very important.” This is illustrated in the Figure 3, which shows the good match between the shape of the intrusion with the resulting ground deformation.

Fig. 3. Correlation between the shape of the ground deformation and the underlying intrusion in a characteristic experiment. a. Topographic map of model surface before the eruption of the oil. b. Plots of the temporal evolution of topographic profiles across the ground deformation pattern. The profiles are located on the map described in a. Each curve of the plots represents a transient stage of the model surfaces during each experiment. One can notice an initial symmetrical ground deformation, evolving to asymmetrical. c. Topographic map of the top surface of the excavated intrusion. The grey scale indicates the depth in mm below the initial surface. The locations of topographic profiles are the same as on the map of model surface in a, such that the profiles of the ground deformation and of the intrusions can be compared. d. Plots of topographic profiles of the model surface (top) and the underlying intrusion (down) parallel to the X-axis (X-profile). e. Plots of topographic profiles of the model surface (top) and the underlying intrusion (down) parallel to the Y-axis (Y-profile). Notice the vertical scale dilation for the profiles of the model surface. The correlation between e and d show that the ground deformation pattern reflects the shape of the underlying intrusion. This suggests that the asymmetrical development of the ground deformation pattern can be analysed to predict where the magma rises towards the surface.

However, Dr. Galland doesn’t spend all of his time in the laboratory. He believes that it is important for him to regularly go into the field to directly observe volcanoes, modern and ancient. He is particularly interested in observing subvolcanic systems, such as the saucer-shaped sills exposed in South Africa’s Karoo region and intrusive complexes formed in compressional settings exposed in the Neuquén Basin in the north Patagonian Andes. Dr. Galland explains, “My approach is to go to the field and really observe and do some detailed work and then out of that comes a question and then these questions can be subsequently addressed with experiments.” After all, that field knowledge is important for turning silica flour into the Earth’s crust and vegetable oil into magma.

Galland, O. 2012. Experimental modelling of ground deformation associated with shallow magma intrusions. Earth and Planetary Science Letters, Vol. 317-318: 145-156.

***Note: Thanks very much to Dr. Galland for providing the three figures and captions.***