2019 Theses Doctoral

The Aleutian arc through and through: Subduction dynamics and the generation, storage, and eruption of hydrous magmas

Rasmussen, Daniel J.

Volcanic arcs are the primary seat of subaerial volcanism and where continental crust is created. Since the advent of plate tectonics theory in the last half century, many of the processes that govern arc magmatism have been described in detail. However, understanding the development and eruption of upper crustal magma reservoirs remains a fundamental challenge. Here we develop and implement new geochemical approaches, and combine our results with those from other disciplines, to explore the location, formation, and eruption of upper crustal magma reservoirs, with the ultimate goal of linking these processes to the underlying process of plate tectonics. Our study area is the central-eastern Aleutian arc, one of the most volcanically active regions on earth, where significant along-strike variability in subduction parameters, magma compositions, and volcanic activity exists.

To advance understanding of magma reservoirs, it is essential to hone our tools for gauging magma depth. Melt inclusion analysis is the first in the toolbox for petrologists, but recent studies have raised questions about the accuracy of this approach. Vapor bubbles commonly form in melt inclusions after entrapment. These bubbles may sequester a substantial portion of the total volatile contents of the melt inclusion, which is problematic because depth estimates are based on melt volatile contents. In Chapter 1, we explore vapor bubble growth in melt inclusions by describing the processes, and their timescales, that lead to bubble growth and developing new methods to retrieve accurate depth estimates from melt inclusions. Our new methods have situational strengths. In concert, they enable extraction of reliable depth information, unlocking the true potential of melt inclusions to measure depth.

With an improved understanding of melt inclusions, we next investigate eruption run-up. During run-up, crustal-scale magmatic systems can be activated, providing a unique opportunity to peer into their structure. In Chapter 2, our goal is to study eruption run-up and determine how magmas are stored in the months, days, and hours leading to volcanic eruption. As a case study, we investigate the 1999 eruption of Shishaldin volcano, of interest because the run-up was months, an unusually long duration, and, despite 43 million cubic meters of tephra ejected in the eruption, no eruption-related deformation was detected in satellite imagery. We develop a new approach for studying run-up that combines diffusion modeling, which gives information on the timing of magmatic processes preceding eruption, with melt inclusion analysis, which gives depth information. Results are combined with those from shear-wave splitting analysis and other geophysical methods. We identify a shallow magmatic system that existed prior to the run-up to the 1999 eruption. A substantial fraction of the magma that was erupted was delivered to the shallow reservoir ~50 days prior to the eruption. More broadly, our results indicate that open-system volcanoes, such as Shishaldin, may commonly have long run-up durations.

Our work on run-up demonstrates the strength of the forensic approach for studying magma reservoirs, but it leads us to question what can be understood in real time. One powerful approach for understanding the state and location of magma reservoirs in real time is the study of volcanic gas emissions. However, interpretation of gas data is a major challenge. To improve our ability to use gas data to understand plumbing systems, and to investigate the shallow magmatic plumbing system of an open-vent volcano, we perform a melt inclusion study of the degassing system at Cleveland in Chapter 3. We focus on Cleveland volcano, one of the most active volcanoes in the US. We develop an empirical degassing model based on melt inclusion data. We use the degassing model to interpret gas composition and flux measurements at Cleveland. Our results indicate gas emissions are generated in a shallow, convecting magmatic system, which is consistent with geophysical observations.

After detailing plumbing systems at Shishaldin and Cleveland, we investigate global trends in magma storage depth in Chapter 4. Geophysically imaged magma storage depths are mostly ~0-20 km depths. The reason for the dramatic variability is not well known. We compile geophysical estimates of magma storage depth and compare these data to magmatic water contents. The initial water content of magma is thought to exert a key control over magma storage depth because as magmas ascent, they degas water. Concurrently, melt viscosity increases and crystallization may be induced. Both these processes promote slowing of magma ascent. We find a strong correlation between magma storage depth and magmatic water contents at the 24 volcanoes that have estimates for both. A global compilation of magma storage depths at 97 volcanoes has a distinct mode at 6 ±3 km, which coincides very closely with the average depth at which arc magmas become water saturated (6 ±3 km) based on maximum water content estimates from a compilation of 77 volcanoes. Melt inclusions from the eastern-central Aleutians do not show evidence of degassing or diffusive loss of water, indicating that water content is a strong control over magma storage depth. The trend exists globally, despite a large range in potential upper crustal controls (thickness, age, stress state, etc.).

In Chapter 5, we move deeper in the arc to understand the underlying processes of subduction and arc magma genesis. Slab depth in the central-eastern Aleutians varies from a near global minimum of 65 km in west (near Seguam) to a more common value of 100 km in the east (near Shishaldin). The cause for this variability is not well known. The thermal structure of the wedge is thought to play a key role in determining where mantle melting occurs, and subduction parameters (slab age, dip, velocity, etc.) exert first order controls on wedge thermal structure. Therefore, subduction parameters are likely to some extent modulate slab depth. However, the mantle-slab coupling depth also has a key influence on the thermal structure of the wedge. Therefore, the coupling depth must also play a role in determining slab depth. A potential third factor is the extent of lateral melt migration in the wedge. In the central-eastern Aleutians, variations in slab depth are reflected in variations in major, trace, and volatile elements. Chemical trends are most consistent with a shallow slab-mantle coupling depth of 50 km throughout the corridor. Results from slab top thermometry suggest that significant lateral variation is unlikely. We speculate that the change in slab depth across the corridor is likely a result of the significant decrease in trench fill sediment thickness moving east.