The 1912 Plinian eruption of Novarupta volcano, Alaska is one the largest historical explosive eruptions and the subject of this study. Plinian eruptions are of very high explosive intensity and destructiveness. They can last for days with individual Plinian phases lasting for hours and producing tens of kilometers tall, sustained columns of hot gases and volcanic particles. These volcanic plumes can become dispersed on a hemispherical or global scale with significant societal impact. Key mechanisms to producing a Plinian eruption are the 'explosive' release of magmatic volatiles, such as water, carbon dioxide and sulfur, in conjunction with the 'explosive' fragmentation of the erupting magma into pyroclasts. The latter are variably sized particles of quenched magma, including volcanic ash. The interrelation between this explosive release of magmatic volatiles and magma fragmentation to result in a Plinian eruption is the focus of this project.
This project will involve the numerical modeling of vesicle size distributions measured in pycroclasts from the 1912 Novarupta eruption. Vesicle size distributions provide a direct record of magma ascent conditions that can be investigated through modeling of bubble nucleation and growth in the ascending magma, both a consequence of volatile exsolution from the magma during ascent-driven decompression. The time-pressure history of an ascending parcel of magma, from which a given pyroclast is derived, is the key parameter determining the fit of modeled to observed vesicle size distributions. This investigation will test the hypothesis that decompression rates are surprisingly large during Plinian eruptions, as well as the question of how bubble nucleation and magma fragmentation are related. The large decompression rates apparently required to nucleate the large numbers of bubbles preserved as vesicles in pyroclasts can only be sustained for very short times, presumably immediately prior to or perhaps during magma fragmentation. It is thought that fragmentation is the consequence of accumulated bubble overpressure in excess of the magma's tensile strength. This project will integrate existing observational analyses of eruptive products with theoretical and empirical work on bubble nucleation and magma fragmentation through a comprehensive quantitative analysis. The resultant reconstruction of magma decompression, bubble nucleation, bubble growth and ultimately magma fragmentation, as recorded by the well-documented 1912 Novarupta pyroclast samples, is expected to lead to new advances in our understanding of explosive volcanism in general and Plinian style volcanic eruptions in particular.
Styles of volcanic eruptions range from effusive to highly explosive. Because of their tremendous destructiveness, the latter pose a considerable threat to human life and the environment, in addition to the potential for significant adverse economic impacts. The objective of this project has been to reconstruct the conditions of magma ascent during Plinian volcanic eruptions. These are sustained explosive eruptions that can last for days, producing columns of volcanic ash that rise for tens of miles into the Earth's atmosphere, where they spread over large distances. The focus of the project has been on two such eruptions, serving as exemplary test cases. One is the largest explosive eruption of the 20th century, the 1912 eruption of Novarupta, Alaska. The other is the 1060 eruption of Medicine Lake volcano, California, an active volcano with potential for future eruptions. Volcanic gases are of essence to explosive volcanism. Water is typically the most abundant volcanic gas and other gases include carbon dioxide, as well as sulfur. Prior to eruption these gases are dissolved in the magma at high pressure. They form gaseous bubbles when magma rises toward the surface and pressure decreases. This imparts buoyancy to the magma, as well as potential energy, due to the high compressibility of gases. During rapid magma ascent bubbles cannot grow fast enough and become pressurized. Ultimately this tears the magma apart through brittle fragmentation, in a manner that is akin to breaking glass. As a consequence, the compressed gases expand explosively, leading to sustained explosive eruptions. Thus, key ingredients for explosive volcanism are the formation of bubbles containing volcanic gases, as well as magma fragmentation. We analyzed quenched magma fragments, called pyroclasts, and discovered that the glassy matrix of these pyroclasts contains substantial amounts of dissolved water (Figure 1). However, we find from thermo-gravimetric analysis and numerical modeling that most of this water is meteoric in origin and must have diffused into the glass during the time after the eruption. In contrast, the amount of magmatic water that remains dissolved in the pyroclasts is very small, indicating that most magmatic water and other volatiles escaped from the magma upon eruption (Figure 2). When we model this syneruptive magma degassing we find that we can reproduce the sizes and abundances of bubbles that formed within the magma, because they are now preserved as vesicles within the pyroclasts (Figure 3). Interestingly, only about half of the magmatic water had formed bubbles by the time the magma fragmented at hundreds of meters depth within the volcanic conduit. The remaining water was therefore lost after the magma had fragmented. From measuring the abundance and interconnectivity of vesicles, we can establish that most of the volcanic gas escaped from the pyroclasts by streaming through interconnect bubbles into newly formed fractures (Figure 3), a process called permeable outgassing. Conventionally it has been thought such outgassing dissipates overpressure, thereby making magma fragmentation less likely. Instead, we are able to show that the interconnected pathways that are formed by bubble coalescence allow the gas to escape rapidly from the pyroclasts as soon as fractures begin to form. As a consequence, the highly viscous magma is rapidly transformed to a dilute gas-pyroclast mixture of substantially reduced viscosity (Figure 4), enabling the magma to erupt at the exceedingly high discharge rates characteristic of Plinian eruptions. Numerical modeling of bubble nucleation and growth indicates that most bubbles form within a few seconds before the magma fragments (Figure 5), whereas scaled laboratory fluid dynamics experiments provide constraints on the rate at which bubbles coalesce (Figure 6). When both are taken into consideration, the picture that emerges is that bubble coalescence does not lead to interconnected bubble networks that are amenable to rapid outgassing of fragmented magma until bubbles reach a volume fraction of greater than 60% (Figure 3). At this stage the magma is also viscous enough for sufficient built-up of overpressure to cause fragmentation. Therefore, it appears that the two necessary conditions for explosive volcanic eruptions of Plinian intensity - sufficient built-up of overpressure for fragmentation and the ability of rapid outgassing upon fragmentation - are both met at a volume fraction of bubbles of greater than 60%. In conclusion, during Plinian volcanic eruptions sufficient overpressure for magma fragmentation is reached on average at bubble volume fractions of greater than 60%. This is also the threshold where the fragmented magma can rapidly release the compressed volcanic gases, in order to be transformed into to a dilute gas-pyroclast mixture of low viscosity, which allows the magma to erupt at high rates. A substantial fraction of the water, which is dissolved in the magma at depth, is released after fragmentation. After deposition meteoric water diffuses over time into the pyroclastic glass, accounting for the high water concentrations measured in a wide variety of pyroclasts.
