[S-IT16] Structure and Dynamics of Earth and Planetary Mantles

[SIT16-09] Combining P- and S-wave Tomographic Filter to Resolve Lower Mantle Compositional Variations

*Jun Su^1,2, John W Hernlund^1,2, Christine Houser¹ (1.Earth-Life Science Institute, 2.Tokyo Institute of Technology)

Keywords:Seismic Tomography, Lower Mantle, Mantle Convection, Geostatistics, Chemical Heterogeneity

The discrepancy between pattern and amplitude of P- and S-wave velocity anomalies in the lowermost mantle (especially below ~2200 km) observed by global seismic tomography is considered evidence of phase transitions or/and compositional differences in denser elements like iron. Alternative explanations for the discrepancy include the influence of decreasing P-wave coverage near the core-mantle boundary and scenarios which invoke only temperature variations. Mantle dynamics models can predict the flow and distribution of these proposed thermal and composition variations, but those models have ideal resolution and cannot be directly applied to the statistical comparison. Therefore, the resolution operator is necessary to bridge the observational and modeling results.

We apply the Singular Value Decomposition (SVD) to obtain the resolution operator to avoid a predetermined global smoothing factor and make the dynamic model comparable to seismic tomography. The resolution operator generated by sensitivity kernels (ray path coverage) acts as a tomographic filter to map the true/theoretical earth structures into the tomographic observation with a corresponding resolution. We implement a statistic analysis to investigate the magnitude distribution of velocity anomaly in lower mantle model shells. The distribution of slow velocity near the core-mantle boundary in filtered dynamic models from Deschamps and Li (2019) shows a similar pattern to the previous study, indicating that lowermost mantle structures are resolvable in both HMSL-P06 and HMSL-S06 models, supporting the existence of compositional/phasic changes.