Okmok volcano in Alaska appears to erupt every decade or so, most recently in 1997 and again in 2008. Its activity has been monitored using modern methods, forming a rich observational data set. The investigators are developing a geophysical model that should explain the timing and location of most of the observed activity. The new information from this study will be directly relevant to understanding time-dependent volcanic hazard posed by Okmok and other, similar volcanoes lying beneath the heavily-used north Pacific air traffic corridors. The same type of model should be applicable to other volcanoes displaying cyclic activity, such as Westdahl in Alaska or Hekla in Iceland.
Okmok is an excellent natural laboratory for such an experiment because a complete cycle of deformation has been monitored using geodetic and seismic means. Using this rich observational data set and a formal protocol for numerical modeling, the investigators are studying Okmok volcano to address the following questions: (1) What is the distribution of material properties within Okmok? (2) How does anelastic rheology modify the temporal evolution of the deformation field? (3) How do material properties at depth influence the deformation observed at the surface? (4) What are the uncertainties of the model parameters that describe the magma chamber?
The impulse-response rheological experiment will improve understanding of the volcano deformation cycle. Specifically, the research project will test four hypotheses: (I) Deformation following the 1997 eruption did not reach a steady state before the eruption in 2008. (II) Viscoelastic stress relaxation contributes to the transient deformation observed during the co-and post-eruptive time intervals. (III) The effective viscosity is several orders of magnitude smaller in the rind of the magma chamber than in the surrounding crust. (IV) The lava flow extruded from Cone A during the 1997 eruption produces a stress field that favors dike propagation from the magma chamber to Cone D.
The results will be published in the international, peer-reviewed literature. A graduate student from a group under-represented in Science, Technology, Engineering and Mathematics (STEM) will be trained for a career in geophysical research at the intersection of three disciplines: seismology, geodesy and volcanology. The modeling protocol and approach will be disseminated among the scientific community at an operational level. Investigator Masterlark, a junior faculty member from an institution funded by the Office of Experimental Program to Stimulate Competitive Research (EPSCoR), will offer a short course on how to apply the Finite Element Method to volcanic deformation. Applicants from underrepresented groups will be especially encouraged to participate. The project will enhance collaboration between the investigators and scientists at the U.S. Geological Survey
The migration of magma within a volcano produces a deformation signature at the Earth’s surface. The internal structure, magmatic characteristics, and stress conditions of a volcano contribute to the specific deformation that can be observed with geodetic techniques. Numerical models can simulate the behavior of these volcanic systems. Forward models can calculate surface deformation for a given scenario of magma intrusion at depth. Unfortunately, such magma intrusion events cannot be observed directly because they are inside the volcanic edifice. Consequently, it is necessary to solve the more challenging problem of developing inverse models that use observed deformation to estimate a few parameters that describe the unknown characteristics of the magmatic intrusion, such as geometry, position, and strength. An accurate understanding of such magma intrusion characteristics is essential to understanding the magma supply of an active volcano, which in turn strongly controls the timing, style, and magnitude of eruption activity. The accuracy of these estimated magmatic intrusion characteristics is controlled by the ability to: (1) resolve the internal structure of the volcano; (2) simulate the magmatic intrusion over a model domain having this internal structure; and (3) estimate the specific parameters that characterize magmatic intrusion from observed deformation. Optimizing the model accuracy requires inverse models that estimate magmatic intrusion parameters, while simultaneously accounting for the geometry and internal structure of a volcano. Prior to this project, no such inverse models with these capabilities were available. The field of seismic tomography successfully addressed topic (1). Seismic analyses have been carried out in support of this geo-mechanical modeling. Initial seismic velocity models were constructed separately from body waves (P- and S-wave velocity, although the latter was poorly resolved) and ambient noise (S-wave velocity only). The seismic study also includes a joint tomographic inversion of the body-wave and ambient noise data. Enhancements to the results from body waves and ambient noise have been achieved via high-accuracy automated S-wave arrival picking and the use of phase-weighted stacking to increase the signal-to-noise ratio of the noise correlation functions. To further constrain the modeling, the project analyzed the deformation of Okmok with synthetic aperture radar images collected by the ERS and Envisat satellites on more than 100 distinct occasions between 1993 and 2008. This time interval encompasses both the 1997 and 2008 eruptions of Okmok. The results show changes in inter-eruption inflation rates that can be interpreted in terms of a Maxwell viscoelastic rheology. From the standpoint of Intellectual Merit, the primary goal of this project was to develop inverse methods that satisfy the requirements for topics (2) and (3). This project successfully developed such methods and demonstrated the implications using the specific example of Okmok Volcano, Alaska. The results indicate that using inverse numerical models of surface deformation to simulate the actual internal structure (as determined from seismic tomography) yield estimates of the magma chamber’s depth that are significantly deeper than the results determined from standard models that ignore the complex internal structure. To account for rheologic complexity, the research project used the Finite Element Method to simulate a pressurized cavity embedded in a viscoelastic medium with material properties derived from body wave seismic tomography and an assumed radial temperature profile. This approach addresses the issue of unreasonably large pressure values implied by a Mogi source in a purely elastic medium. Assuming reasonable values for the Maxwell viscosity, the viscoelastic model can explain the exponentially decaying relaxation observed in the five years following the 1997 eruption. In summary, these more realistic numerical models provide more reliable descriptions of the interior process of active volcanoes, than do standard models that oversimplify the interior structure of a volcano. In terms of broader impacts, the inclusion of the more complex and realistic internal structure in the numerical models significantly improves their usefulness in assessing hazard. The project trained a female graduate student who successfully completed the requirements for a Master's degree in Geophysics at the University of Wisconsin-Madison in 2010. Now a candidate for a Ph.D. at the University of Wisconsin-Madison, she plans to defend her Ph.D. thesis in 2014. The project offered a three-day short course on modeling volcano deformation at the UNAVCO facility in Boulder Colorado, May 21-23, 2013. More than 20 individuals applied for the short course. Of these, eleven graduate students pursuing M.S. or Ph.D. degrees at U.S. institutions were selected. The participants included students from historically under-represented groups. This project has also enhanced the collaboration between the scientists at the Universities of Wisconsin and Alabama and scientists at the U.S. Geological Survey’s Alaska Volcano Observatory.
