Research projects

The Alaska-Aleutians subduction zone is one of the most seismically active regions of the world, having hosted the second-biggest earthquake recorded in history. However, if we look closely, certain sections of the subduction present a lot of big earthquakes and tsunamis but some others are much more quiet. 

Experiment

We identified two side-by-side sections with this contrasting earthquake behavior within the Andreanof segment of the Aleutians Islands: Adak and Atka. GPS data show that Adak, the area with a higher recurrence of major earthquakes, is presently locked, signifying that the two tectonic plates are stuck together, accumulating energy for future earthquakes. In contrast, the Atka section experiences creep, involving slow movement without energy accumulation.

So we went there! we collected seismic reflection and refraction data in search of answers (Well, actually we were more interested in the arc formation and magma transfer. You can read more about it here and see a comic about my life onboard the research vessel Marcus G. Langseth here.)

Preliminary Results

We found that the incoming Pacific plate looks the same in both sections, meaning that changes in its topography roughness, sediment thickness or composition, and/or water budget could not account for the varying seismic behavior. So we turned to the overriding plate and found a significant difference in deformation, with the locked section (Adak) showing more shortening or compression. Bottom-simulating reflectors (BSR) are commonly used to constrain thermal structure in subduction zones. The BSR is shallower in the Adak forearc than the Atka forearc, suggesting higher shallow geothermal gradients there, which we interpret as heat advection due to more efficient dewatering of the forearc. Thus, we propose that Adak has a better-developed fault network that may enable efficient fluid migration, lowering pore fluid pressure and facilitating coupling at the megathrust. 

Like in many other subduction zone systems, in south Chile the obliquity of the subduction transfers stress to the continent that is partitioned as compression and shear stress. The shear stress is parallel to the trench and creates strike-slip systems in the overriding plate. In south Chile, this strike-slip system is known as the Liquiñe-Ofqui fault zone. 

One problem in mapping the Liquiñe-Ofqui fault zone is that south Chile has a lot of vegetation, rainy weather, and sections underwater, so geophysics comes in handy. Our efforts are currently focused on interpreting seismic reflection data collected in the Gulf of Ancud in 2002.

Our preliminary observations reveal deformation within the fluvial and glaciomarine sediments that compose the shallow layers of the basin. The trace of the Liquiñe-Ofqui fault shows a negative flower structure, encompassed by extensional deformation evident through normal faulting. These faults extend to the surface and look fresh, indicating recent activity and an ongoing active system.

(This is my CIDER project!)

When I began studying earthquakes, I only knew about the stick-slip model, and that was it. Now, we know that fault slip occurs across a broad range of rates, encompassing regular earthquakes to slow slip events (SSEs). Researchers have developed computational earthquake cycle models to investigate slip rates along the megathrust. They manage to replicate slow slip by incorporating heterogeneous friction and low effective normal stress due to high pore fluid pressure.

Questions

We recognized that factors such as temperature and lithology dependence of frictional properties, as well as their distribution, remain underexplored in SSEs modeling.

• Can we model the recurrence and size of SSEs by changing the lithology and distribution of heterogeneities? 

• What if we can replicate slow-slip through adjustments in frictional properties and distribution, without necessarily requiring high pore fluid pressure? and, if it turns out that we do indeed need pore pressure, do we truly need the high levels that other models employ?

Approach

Using the code HBI from Ozawa et al., 2023, we simplify the subduction interphase into a matrix with patches of different lithology. Based on the exhumed subduction rock record and laboratory experiments, we constrained the interface's structure, lithology, and temperature dependence of frictional properties.

For our study, we are focusing on the Cascadia margin. We plan to vary possible interface patch/matrix lithologies and compare model results to existing seismic and geodetic data to assess the reproducibility of SSE spatial and temporal variability using this heterogeneous interface. 

We expect to have results by the end of the year. My collaborators will be presenting this work as a poster at the AGU fall meeting, 2023.