Joint Interpretation of Global Shear Attenuation and Wave-speed models in terms of temperature, melt content and composition of the mantle – JIGSAW2

Seismology has been very successful in mapping the 3D distribution of seismic velocities within the Earth mantle. However, despite tremendous progress, first order fundamental questions remain unanswered. These include the origin of the low velocity zones observed at the bottom of the mantle, near the transition zone or below the lithosphere, the amount of melt and water present in the mantle, and the origin of the long-term stability of cratons. To answer these questions, we need to obtain robust constraints on physical properties of the Earth such as temperature, composition, volatile content, or presence of partial melt. This is necessary to make direct links with mantle convection and geochemistry. A strong limitation is that seismic velocity alone is not sufficient to constrain these physical properties and complementary observations with different sensitivities to these quantities are required. Therefore, mapping attenuation (the loss of energy of seismic waves) in addition to elastic velocity or anisotropy (the directional variation of seismic velocities, a proxy of mantle flow) is critical.

The JIGSAW2 project aims at mapping 3D variations in attenuation and velocity, and at interpreting them simultaneously. The JIGSAW2 team consists of two partners (Lyon and Canberra) with internationally recognized researchers strongly motivated to tackle this challenge. Recent numerical techniques for wavefield computations permit indeed further advances in attenuation tomography. Only two groups (Princeton and Berkeley) have exploited these numerical approaches for mapping shear attenuation at continental and global scales and a few other teams have published global 3D models of shear attenuation using more standard approaches. All these models are obtained using linearized inversion approaches, which only allows small and smooth deviations from an average 1D depth-dependent seismic reference model.

In JIGSAW2, we will build a new set of attenuation and travel-time measurements using numerical 3D synthetics that fully account for 3D heterogeneities. We have developed an approach to obtain accurate measurements of long period S body waves and Rayleigh surface waves from the comparison of seismic records with 3D synthetics. We will extend this approach to Love waves for extracting radial anisotropy, and long period P-waves for imaging the transition zone and lower mantle. We expect to obtain a global database of P-wave velocity, and of S-wave attenuation and velocity measurements in the Earth’s mantle.

These databases will then be jointly inverted to obtain a 3D model of P wave speed, S wave speed and attenuation, taking into account radial anisotropy. We will use a trans-dimensional tree Bayesian inversion framework recently developed by the Canberra group. This approach is more efficient than previous trans-dimensional algorithms and we have already applied it to global tomography with a subset of our current dataset. Compared with standard linearized inversion techniques, it does not require model regularization (smoothing). It will therefore more accurately retrieve the amplitude of variations and allow sharper velocity gradients in the model. It will also provide a thorough probabilistic estimate of uncertainties. These advantages are keys to interpret seismic models in terms of physical parameters.

We will jointly invert the shear attenuation and the seismic velocities, to constrain temperature, melt content and composition of the mantle. Again, we will use Bayesian approaches in order to quantify trade-offs and uncertainties.

We expect that JIGSAW2 will help to better understand the origin of seismic heterogeneities and the structure and dynamics of the Earth’s interior. The scientific breakthroughs will come from the use of full 3D synthetics, the large focus on attenuation, and the use of advanced probabilistic approaches.