The Earth’s mantle consists of solid rocks and also viscous mantle flow. The process that drives the mantle flow between Earth’s surface and deep interior is known as mantle convection. The mantle flow appears as mantle plume or subducting slab that is the driving force of the plate motions on Earth’s surface. The mantle transition zone is a 250-km thick layer located in the middle of mantle and acts like a filter in convection between the upper and lower mantle. The mantle flow in the transition zone is the key to understanding whole mantle convection. How can we trace the mantle flow in deep Earth’s interior, especially in the transition zone? Mantle flow can cause the preferred orientation of minerals, thus produce seismic anisotropy that can be observed from seismic waves (Figure 1). Seismic anisotropy, which is the dependence of seismic velocities on the propagation direction or seismic wave polarization, is a powerful tool to constrain the nature of mantle flow. However, current seismology methods like surface wave or shear wave splitting have limited resolution to observe seismic anisotropy in the transition zone. Here, I propose to use a body wave method, SS precursors, to observe and quantify the strength of seismic anisotropy in the transition zone. The SS precursor method has a better vertical and horizontal resolution because it can pinpoint the seismic anisotropy in the transition zone beneath a specific location and lead to a better understanding of flow in the deep Earth.
Figure 1. The minerals in mantle
transition zone are aligned by the subducting flow.
Figure 2. The ray-path of SS phase and SS precursors (Schmerr et al., 2010).
2. Methods
I’m currently working with a global broad band SS dataset consisting of 45,624 records (Figure 3). I partitioned the SS dataset into different geographical bins located in South America and South Pacific oceans, which have enough azimuthal coverage to produce stable stacking results. I also further broke the geographic bins into smaller azimuthal bins to study the travel time and amplitude variance of SS precursors with azimuth. If seismic anisotropy exists in the transition zone, it can change the amplitude and travel time of SS precursors at different azimuths. The azimuthal stacking results show weak variations of amplitude and travel time, which is consistent with less than 1% anisotropy beneath South Pacific Ocean and South America. In order to trace subducting flows in the mantle, I combined all the geographical bins in subduction zones into a large bin and then broke it into 6 different azimuthal bins (Figure 4). The variations of travel time and amplitude are relatively strong compared to the bins in South America and South Pacific Ocean, which indicates 1-4% anisotropy can exist in the transition zone beneath subduction zones (Figure 5).
Figure 3. The azimuths of SS bounce
points in the SS dataset.
Figure 4. The stacking results of S410S
and corresponding synthetics.
Figure 5. The amplitude and travel time variations of S410S with azimuth.
3. Proposal: 3D Modeling of SS Precursors
I propose to use the 3D SPECFEM code to model
the amplitude and travel time change of SS precursors in an anisotropic 3D
Earth model. I will learn to use SPECFEM code to generate synthetic seismograms
and predict how the waveforms of SS precursor would behave if the transition
zone has certain amount of anisotropy. The modeling results will be used to
quantify the strength of anisotropy observed in my dataset by examining the travel
time and amplitude variations. First, I will set up the SPECFEM code for
running on the Deepthought2 cluster. Since it is my first time to use the code,
it will take me about two weeks to study the tutorial and adapt the code to
work on the cluster. Second, I will collect anisotropic Earth models from
literatures and use them as an input to test the code. If the synthetics are
consistent with data, I can continue to adapt the models to test certain
hypotheses. Third, I will specify different strength and azimuth of seismic
anisotropy in the transition zone. To
begin with, I will test the resolution of SS precursor method for the minimum
anisotropy that can be detected by the method. I will incrementally add 0.1% of
azimuthal anisotropy into the model until the amplitude and travel time changes
are observed. From this I can determine the minimum anisotropy to be seen in
the dataset. Furthermore, I will run models with different strength of
anisotropy ranging from 1 to 15% to find the reasonable anisotropy that can
explain the data variations. I can fit the model predictions into the data
observations by creating sinusoidal curves with different strengths of
anisotropy. The goal is to find the best-fit model thus to quantify the amount
of anisotropy in the transition zone.
Figure 6. Shear wave velocity anomalies from s20rts model on the SPECFEM mesh.
Figure 7. An example to use SPECFEM for mantle convection modeling.
References
Shearer, P. M. (1993). Global mapping of upper mantle reflectors
from long-period SS precursors. Geophysical Journal International, 115(3),
878–904.
Flanagan, M. P., & Shearer, P. M. (1998). Global mapping of topography
on transition zone velocity discontinuities by stacking SS precursors. Journal
of Geophysical Research: Solid Earth, 103(B2), 2673–2692.
Mainprice, D. (2007). Seismic anisotropy of the deep Earth from a
mineral and rock 1022 physics perspective. Schubert, G. Treatise in Geophysics
Volume 2 pp437-492.
Schmerr, N., & Garnero, E. J. (2007). Upper Mantle Discontinuity
Topography from Thermal and Chemical Heterogeneity. Science, 318(5850),
623–626.
Schmerr, N., & Garnero, E. (2006). Investigation of upper mantle
discontinuity structure beneath the central Pacific using SS precursors. Journal
of Geophysical Research: Solid Earth, 111(B8), B08305.
Trampert, J., & van Heijst, H. J. (2002). Global Azimuthal
Anisotropy in the Transition Zone. Science, 296(5571), 1297–1299.
http://doi.org/10.1126/science.1070264
Yuan, K., & Beghein, C. (2013). Seismic anisotropy changes
across upper mantle phase transitions. Earth and Planetary Science Letters,
374, 132–144. http://doi.org/10.1016/j.epsl.2013.05.031
