Testing the role of metastable olivine in subduction dynamics and deep earthquakes

University Of California-Davis, Davis CA

Investigators

Abstract

Plate tectonics on the Earth’s surface is driven by convective motion of solid rocks the mantle. Mantle thermal convection is due to density differences arising from differences in chemical composition, mineral structure (phase), and temperature. The primary driving force for convection arises from gravitational sinking of cold, and therefore dense, tectonic plates within subduction zones. It has previously been proposed that within the sinking tectonic plate (called subducting slab), the cold temperature can delay the normal changes in mineral structure that occur with increasing pressure. This causes the plate to be less dense than expected based on temperature alone. It has also been proposed that this less dense region - called a metastable olivine wedge - is necessary to trigger earthquakes that occur at depths of > 400 km. These earthquakes are referred to as deep earthquakes. However, it is still unknown whether metastable olivine exists in all slabs where deep earthquakes are observed. There are also other possible mechanisms that may trigger deep earthquakes. Here, the researchers use state-of-the-art numerical simulations to test how the presence of metastable olivine affects subducting-slab dynamics. They integrate in the modeling various experimental and observational parameters and test different scenarios. They compare the model outputs with observations in existing subducting slabs around the Globe. Results of this study directly inform potential origins of the different slab dynamics and that of deep earthquakes. The new improved computational codes are shared with a scientific community. The project has strong implications for the understanding the dynamics of subduction zones, at the origin of the largest earthquakes threatening human societies. The project also provides support and training for several graduate and undergraduate students at University of California - Davis. One of the biggest outstanding questions of mantle dynamics is why some slabs appear to stagnate in or below the transition zone, while others appear to sink directly into the lower mantle. The answer is related to phase transitions and trench motion, but it is unclear how these feedback with other material properties and larger-scale mantle flow to generate the apparent variability in slab behavior. This project tests the role of a metastable olivine wedge (MOW) in subduction dynamics and its potential to serve as a source for deep earthquakes. The team first overcome several simplifications of previous models through integration of HeFESTo for equilibrium phase transitions - including density and latent-heat effects, and a grain-size and water-dependent metastable olivine transformation (MOT) model- into a fully dynamic subduction simulations with a visco-plastic rheology including Peierls creep. The simulations run in the software Aspect. 2D models test how slabs with different ages and rates of subduction (controlled by the plate boundary shear zone; PBSZ) lead to different subduction dynamics. Observations of slab morphology together with plate and trench motion are compared to model output to constrain uncertainty in parameters including PBSZ viscosity, the gradient of viscosity into the lower mantle, and the water content of the slab. In addition, the researchers compare the distribution of the MOW and its overlap with strongly deforming regions of the slab; the aim is to predict the expected pattern of seismicity. This pattern is compared to observations and analyzed accounting for the thermal structure of the slab (e.g., is the pattern indicative of warm versus cold slabs). The team uses 3D models to quantify the impact of trench width together with along-strike variation in subducting plate age and PBSZ properties. While these are generic models, the initial conditions are chosen to represent three different subduction zones: Japan-Izu-Bonin, Tonga-Kermadec and South America. Using the observations of MOW extent in Japan, Izu-Bonin, and the Marianas, the team calibrates the MOT model, and use it to predict the MOW extent for the other two regions. The model results are compared with observations using along-strike variations in the slab morphology, plate motions, and seismicity as a further constraint on model parameters. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.