Petrologic Reconstruction of Magmatic System Variables and Processes

In summary, petrologic studies are so integral to the investigation of magmatic systems that they are often taken for granted by those who apply new geophysical and geochemical monitoring techniques to active volcanoes. Here we have shown ways in which the traditional strengths of petrologic studies of magma generation and storage can be integrated with real-time monitoring data to connect direct investigations of magmatic processes (particularly volatile exsolution and crystallization) with manifestations of those processes that are used in monitoring, such as seismicity and gas emissions. "Traditional" petrologic methods used to study magma storage conditions focus on constraining intensive parameters such as temperature, pressure, oxygen fugacity and dissolved volatile contents. Our review of commonly-used mineral-mineral geothermometers shows that they are generally robust, with Fe-Ti oxides providing the best constraints. In contrast, geobarometry is notoriously difficult, particularly for upper crustal pressures. For this reason, we review alternative methods of estimating equilibration pressures, including phase equilibria experiments, volatile saturation models, matrix glass compositions and crystal textures. As no one method is perfect, we recommend using as many different methods as permitted by the sample set. Mineral-melt thermobarometry (Putirka 2008) offers significant advances in this regard. We confine much of our comparison of petrologic and monitoring data to the well-characterized eruption of Mount St. Helens in the 1980s. This is not to suggest that Mount St. Helens should, in any way, be considered a reference standard, but simply because both the nature of the eruption (with styles ranging from plinian to sub-plinian, vulcanian, and effusive) and the wealth of monitoring and petrologic data allow us to compare these datasets for different eruptive conditions. We find an impressive correlation between the petrology and the monitoring data, which provides some interesting insights into the chemical and physical processes occurring during the 1980-1986 eruption. Specifically, comparison of fo2 recorded in Fe-Ti oxides and crater fumaroles demonstrates that shallow degassing causes a reduction in magmatic fo2, as predicted by Burgisser and Scaillet (2007). Comparison of petrologic and measured SO2 emissions shows these estimates to be similar for eruptive episodes, but implicates non-erupted magma in generating inter-eruptive gas fluxes, thereby raising questions about the physical mechanisms of gas transport in active, but non-erupting, magmatic systems. Comparison of petrologic estimates of magma storage pressures with seismic data provides important insights into the development of subvolcanic regions of magma storage and transport and raises the tantalizing possibility that the geometry of subvolcanic regions may control the style of ensuing eruptions. Where large numbers of melt inclusions from a single eruption are analyzed for H2O and CO2, there exists the possibility of constraining conduit diameter, an essential, but poorly-constrained, parameter in dynamical modeling. Improved geospeedometry techniques (either chemical or textural) would also improve geometric constraints. In stepping back to examine the evolution of magma storage regions between eruptions, we demonstrate the importance of fully characterizing the phase proportions and composition of individual samples, as well as the full range of melt inclusion compositions contained within those phases. Simple models of degassing and crystallization illustrate the potential for comprehensive melt inclusion studies to elucidate magma decompression and crystallization histories, particularly when combined with trace element and isotopic analyzes of the same inclusions. Application of our simple models to examples of both large silicic and small intermediate systems demonstrates the profound difference in these systems, from their geometry and crystallization history to implied conditions of magma input and eruptive frequency. Generalization of the conclusions drawn from examination of the few systems for which complete data sets are available will require generation of similar (or improved, particularly by addition of textural data) datasets for more magmatic systems. As the tools for such analysis are now readily available, we hope this review will inspire collaborative efforts to fully characterize many different types of magmatic systems, and to integrate those observations with studies in physical volcanology, numerical modeling and volcano monitoring. Copyright © Mineralogical Society of America.