Normal mode based seismic elastic and anelastic tomography and its interpretation in terms of mantle temperature, composition and grain size
Abstract
Tectonic processes at Earth's surface, are driven by convection deep in Earth's mantle. Seismic tomography using earthquakes is the main tool to directly image the Earth's mantle. Such images show regions with slow and fast velocity anomalies, but only velocity doe snot provide enough information to interpret these anomalies in terms of temperature, composition and grain size. Additional information such as density and attenuation is required to make these interpretations. Density is a key parameter to determine buoyancy and attenuation links to viscosity and grain size, so together they may constrain mantle convection.
Here, I will show that the key to making tomographic models of velocity, density and attenuation whole Earth oscillations, which are being observed in low frequency seismic spectra. One of the advantages of using whole Earth oscillations is that we can use exact theory to calculate the effect of 3D mantle structure variations on low frequency seismic spectra. Furthermore, these whole Earth oscillations do not only provide shear wave velocity, but also additional information such as density and attenuation and also discontinuity topography. Recent computer advances have now made it possible to incorporate normal modes with ever increasing complexity into seismic tomography, including cross-coupling (or resonance) between groups of modes and directly inverting the frequency-domain normal mode data in a `full-spectrum' tomography, analogous to the `full-waveform tomography which is now common practice in time-domain tomography. We find that the two Large Low Shear Velocity Provinces (LLSVP) in the lower mantle are partially dense at their base while the remaining parts of the LLSVPs are light. Using a recently developed neural network, we interpret our elastic model which shows that the dense and light parts both have a high temperature; the larger density comes from an increase in iron content which may be stable in mantle convection. We also find that the LLSVPs have weak attenuation. Comparing our anelastic model with a laboratory-based viscoelastic model, we interpret the weak attenuation in combination with high temperature as being due to a large grain size, which would increase the viscosity further suggesting that the LLSVPs may be stable.- Publication:
-
AGU Fall Meeting Abstracts
- Pub Date:
- December 2022
- Bibcode:
- 2022AGUFMDI23A..01D