Recent progress in development and application of quantitative and objective inversion for body-w... more Recent progress in development and application of quantitative and objective inversion for body-wave waveforms has allowed us to obtain fine 1-D (vertical) structure for several regions (e.g., D" layer studies, Kawai et al., 2007ab, GRL). The inverse problem is written as: A δ m = δ d_OBS-INIT, [eqn:observation] where A is the matrix of partial derivatives, δ d_OBS-INIT is the residual of observed waveforms and synthetic waveforms for an initial model, and δ m is the perturbation to the initial model. We minimize the residual of the observed and the synthetics to obtain a modified model. Utilizing a dataset of many waveforms taken as a whole, we can extract robust information on the Earth's structure, even though there might be a relatively large difference between the observed data and the initial synthetics. The robustness of the inversion results is confirmed in several ways. In our inversion we use the SVD (singular value decomposition) for the matrix A: A = U Λ VT and VT V = [ δ m1 δ m2 δ m3 ·s ]. We represent the modified model δm as: δ m = M1δ m1 + M2δ m2 + ·s + Mnδ mn. Here we introduce a factor which we call "SVD sensitivity for each datum," which is the inner product of each observed waveform and the modified waveform for an individual model change. Plotting these factors helps us to visualize the distribution of the dataset and to extract robust information on regional structure. As we have mainly studied the mantle transition zone beneath Japan and environs, we find some structure which might represent the cold slab stagnating into the transition zone near Hokkaido, or lower Qμ for all the regions of our interest than the global average PREM model, which might be connected to thermal or compositional structure there.
We invert shear-wave waveform data for the radial variation of (isotropic) shear-velocity in D″ b... more We invert shear-wave waveform data for the radial variation of (isotropic) shear-velocity in D″ beneath Northern Asia. We reduce source and receiver effects by using data for intermediate and deep events beneath Italy and Japan recorded respectively at stations in East Asia and Europe. Relative to PREM, we find a significantly higher S-wave velocity in the depth range from 150 to 300 km above the core-mantle boundary (CMB) and a slightly lower S-wave velocity in the depth range 0-150 km above the CMB. As our previous studies of D″ structure beneath Central America and the Arctic obtained similar S-wave velocity models, we suggest that this pattern of vertical dependence of shear wave velocity in D″ may be a general phenomenon, at least in relatively cold regions.
Inversion of body-wave waveform data for elastic and anelastic transition-zone structure in and a... more Inversion of body-wave waveform data for elastic and anelastic transition-zone structure in and around Japan Nobuaki Fuji, Kenji Kawai and Robert J. Geller We have developed new methods to invert seismic waveform data for localized seismic structure. We use these methods to invert for the fine structure of the mantle transition zone in and around Japan using broadband data from regional arrays. Our methods use tools we have developed for performing 'static corrections' for the complex crustal structure in and around Japan. We conduct static correction by making a time shift which gives the best correlation coefficient for the first arrival S phase. We also have used other methods for time shifting such as onset pick time shift, but we found that the results of the inversion do not greatly differ. In order to stabilize the inversion we use automated data selection criteria, which limit the amplitude ratio of the synthetic seismogram to the observed seismogram to be between 0.3 and 3.0, and require a value above 0.5 for the correlation coefficient. The amplitude ratios (synthetic/observed) are systematically large for the initial (PREM) anelasticity model, so we invert simultaneously for both Q and elastic structure. We invert the transverse components of long-period (20-200 s) body-wave data from the NIED F-net array. The target regions lie beneath Hokkaido and the Philippines Sea at depths between 200 km and 700 km. The dataset used in this study consists mainly of triplication S phases that sample the mantle transition zone. It is difficult to analyze such data using previous methods, but our methods can handle such data. Models from studies of this type should contribute to improving our knowledge of the thermal state and proportion of water present in the upper mantle.
Using localised body-wave waveform inversion method for both elastic and anelastic structure (whi... more Using localised body-wave waveform inversion method for both elastic and anelastic structure (which we have recently developed), we found relatively high attenuation in the mantle transition zone beneath in and around Japan. We divide the region of interest into several sub-regions and obtained 1-D structure model both for elastic and for anelastic structure. From plural aspects of statistics (stacked waveforms, AIC, data distribution for kernels), we confirm that the high attenuation is the key to realise the amplitudes of the waveforms particularly (Fuji et al., in prep.). In this presentation, we will mention on our methodology briefly and show our results. Beneath the mantle wedge (Philippines Sea), we found higher attenuation than beneath the Pacific Ocean. This will suggest some anomaly information either on the temperature or on the water content. We will discuss the meanings further in the last of the presentation.
73rd EAGE Conference and Exhibition - Workshops 2011, 2011
In the last decade, the deployment of dense regional arrays such as the USArray transportable arr... more In the last decade, the deployment of dense regional arrays such as the USArray transportable array has considerably improved our capacity to image the interior of the Earth. However, the use of the wealth of information coming from these new and large amounts of high quality ...
We are carrying out small scale pilot efforts to invert actual data for the fine structure of D&q... more We are carrying out small scale pilot efforts to invert actual data for the fine structure of D" using waveform inversion. These efforts will then be significantly scaled up. Now, we consider broadband transverse component S-wave waveform data recorded in North America for deep events under South America. These events sample the lowermost several hundred km of the mantle (the D" region), immediately above the core- mantle boundary (CMB) under Central America. Interest in the D" region has recently become even more intense due to the discovery of a high-pressure phase (post-perovskite, abbreviated as "ppv"), which is likely to be responsible for the jump in seismic velocity at the top of D". For the moment our preliminary inversion seeks a laterally homogeneous isotropic S-velocity model appropriate to the base of the mantle under Central America. Since the inversion is only for the structure of D" in the target region, other effects must be account...
We present a method for calculating Fréchet kernels from a precomputed database of strain Green&#... more We present a method for calculating Fréchet kernels from a precomputed database of strain Green's tensors by the Direct Solution Method, which allows us to go to frequencies up to 1 Hz. In order to accelerate the kernel calculations, we interpolate and reconstruct the Green functions for arbitrary epicentral distances from the Green's tensor database. The algorithm is very efficient and allows us to calculate waveform, traveltime, and amplitude sensitivity kernels for high-frequency teleseismic body waves on a laptop. We will show travel-time and amplitude sensitivity kernels for high frequency P, PKP and Pdiff phases. We will also present waveform Fréchet derivatives in the distance range corresponding to the mantle transition zone discontinuities triplication and to PcP precursors. The temporal dependence of these Fréchet derivatives shows a high-degree of complexity, that can be exploited to resolve fine scale structures by waveform inversion.
In the last decade, the deployment of dense arrays such as the USArray transportable array has co... more In the last decade, the deployment of dense arrays such as the USArray transportable array has considerably improved our capacity to image the interior of the Earth. However, the exploitation of the wealth of information coming from these new and large amounts of high quality broadband data still heavily relies on asymptotic approaches. Here we present a new method that couples a spectral element method inside a regional 3-D domain with a global propagation approach in a spherically symmetric Earth model. This method is both very accurate and very efficient, even at periods as short as 1 s. It allows us to model short period teleseismic waves in 3-D media, and to compute 3-D waveform Fréchet derivatives, opening the way to 3-D waveform inversion of teleseismic body wave records from dense, local or regional, broadband arrays.
ABSTRACT Inversion of body-wave waveform data for elastic and anelastic transition-zone structure... more ABSTRACT Inversion of body-wave waveform data for elastic and anelastic transition-zone structure in and around Japan Nobuaki Fuji, Kenji Kawai and Robert J. Geller We have developed new methods to invert seismic waveform data for localized seismic structure. We use these methods to invert for the fine structure of the mantle transition zone in and around Japan using broadband data from regional arrays. Our methods use tools we have developed for performing 'static corrections' for the complex crustal structure in and around Japan. We conduct static correction by making a time shift which gives the best correlation coefficient for the first arrival S phase. We also have used other methods for time shifting such as onset pick time shift, but we found that the results of the inversion do not greatly differ. In order to stabilize the inversion we use automated data selection criteria, which limit the amplitude ratio of the synthetic seismogram to the observed seismogram to be between 0.3 and 3.0, and require a value above 0.5 for the correlation coefficient. The amplitude ratios (synthetic/observed) are systematically large for the initial (PREM) anelasticity model, so we invert simultaneously for both Q and elastic structure. We invert the transverse components of long-period (20-200 s) body-wave data from the NIED F-net array. The target regions lie beneath Hokkaido and the Philippines Sea at depths between 200 km and 700 km. The dataset used in this study consists mainly of triplication S phases that sample the mantle transition zone. It is difficult to analyze such data using previous methods, but our methods can handle such data. Models from studies of this type should contribute to improving our knowledge of the thermal state and proportion of water present in the upper mantle.
By using waveform inversion to study the seismic velocity structure of D'', we found th... more By using waveform inversion to study the seismic velocity structure of D'', we found that the upper half of D'' beneath Central America has S-wave velocity significantly in excess of PREM, while the lower half has S-velocity nearly equal to PREM [Kawai et al., GRL, 2007]. We interpret this as evidence for a double crossing phase transition (a reverse transition from post-perovskite=ppv to perovskite=pv), the existence of which was suggested by Hernlund et al. [2005, Nature]. In addition to the D'' layer beneath Central America, we have now also inverted seismic body- wave waveform data for the vertical dependence of (isotropic) shear-velocity in D'' beneath Asia and the Arctic, also using the transverse components of relatively long period broadband waveforms (20-200 s) as data. We found that these data also suggest the existence of high S-velocity in the upper half of D'' and low velocity in the lower half, consistent with the existence of a double crossing phase transition within D''. Our results for these three separate regions, when taken together, suggest that the existence of a double crossing phase transition is widespread and possibly global or nearly so. This has important implications for studies of the mineral physics, temperature profile, convection and material transport in D''. In the near future, we plan to study D'' beneath Africa and the Pacific in order to further determine the extent to which the double crossing phase transition is (or is not) ubiquitous.
ABSTRACT Recent progress in development and application of quantitative and objective inversion f... more ABSTRACT Recent progress in development and application of quantitative and objective inversion for body-wave waveforms has allowed us to obtain fine 1-D (vertical) structure for several regions (e.g., D" layer studies, Kawai et al., 2007ab, GRL). The inverse problem is written as: A delta m = delta d_OBS-INIT, [eqn:observation] where A is the matrix of partial derivatives, delta d_OBS-INIT is the residual of observed waveforms and synthetic waveforms for an initial model, and delta m is the perturbation to the initial model. We minimize the residual of the observed and the synthetics to obtain a modified model. Utilizing a dataset of many waveforms taken as a whole, we can extract robust information on the Earth's structure, even though there might be a relatively large difference between the observed data and the initial synthetics. The robustness of the inversion results is confirmed in several ways. In our inversion we use the SVD (singular value decomposition) for the matrix A: A = U Lambda VT and VT V = [ delta m1 delta m2 delta m3 ·s ]. We represent the modified model deltam as: delta m = M1delta m1 + M2delta m2 + ·s + Mndelta mn. Here we introduce a factor which we call "SVD sensitivity for each datum," which is the inner product of each observed waveform and the modified waveform for an individual model change. Plotting these factors helps us to visualize the distribution of the dataset and to extract robust information on regional structure. As we have mainly studied the mantle transition zone beneath Japan and environs, we find some structure which might represent the cold slab stagnating into the transition zone near Hokkaido, or lower Qmu for all the regions of our interest than the global average PREM model, which might be connected to thermal or compositional structure there.
ABSTRACT Using localised body-wave waveform inversion method for both elastic and anelastic struc... more ABSTRACT Using localised body-wave waveform inversion method for both elastic and anelastic structure (which we have recently developed), we found relatively high attenuation in the mantle transition zone beneath in and around Japan. We divide the region of interest into several sub-regions and obtained 1-D structure model both for elastic and for anelastic structure. From plural aspects of statistics (stacked waveforms, AIC, data distribution for kernels), we confirm that the high attenuation is the key to realise the amplitudes of the waveforms particularly (Fuji et al., in prep.). In this presentation, we will mention on our methodology briefly and show our results. Beneath the mantle wedge (Philippines Sea), we found higher attenuation than beneath the Pacific Ocean. This will suggest some anomaly information either on the temperature or on the water content. We will discuss the meanings further in the last of the presentation.
ABSTRACT The lowermost mantle (which is called the D'' region) is a boundary laye... more ABSTRACT The lowermost mantle (which is called the D'' region) is a boundary layer for mantle convection and is also a boundary layer between the outer core (composed primarily of molten iron) and the mantle, where compositional differentiation and thermal/chemical interaction can occur. Experiments by Murakami et al. (2004) found a phase transition from perovskite structure to post-perovskite structure at pressures corresponding roughly to the top of D''. Better knowledge of the seismic fine structure of D'' region thus can make an important contribution to understand the dynamics of the lowermost mantle. We have developed and implemented methods for inversion of seismic waveforms for Earth structure, and have applied them to data from dense seismic arrays (e.g., Kawai et al., 2007a,b, GRL). In a recent application of these methods, Konishi et al. (2009, EPSL) found evidence that basaltic crust is subducted to the lowermost mantle, on the basis of the fine S-wave velocity structure of D" beneath the Western Pacific. In this study, we conduct waveform inversion for the fine shear velocity structure in the lowermost mantle beneath Central Asia using data from ORFEUS and IRIS for earthquakes that occurred in 1995-2006 beneath Southeast Asia with epicentral distances of roughly over 75 degree. We use data bandpass filtered from 8-200s to invert for the vertical dependence of the shear velocity. We obtain a model whose the average shear velocity is faster than PREM in D''. We are conducting further inversions to obtain more information on this region using higher frequency data and more detailed regionalization.
ABSTRACT It is impractical to directly constrain the elastic constants of a transversely isotropi... more ABSTRACT It is impractical to directly constrain the elastic constants of a transversely isotropic (TI) medium using travel-time data. In contrast, as we show in this study, the elastic constants can be determined straightforwardly by inversion of body-wave waveform data. We invert the horizontal components of observed seismic waveforms of S and ScS phases (as well as any other phases arriving in the time window) to determine the radial profile of TI shear wave velocity in the lowermost mantle beneath the Pacific. We find that the radial (SV) component is faster than the transverse (SH) component in the depth range from about 200--400~km above the core-mantle boundary (CMB). The major mineralogical components above the D" discontinuity in this depth range are Mg-perovskite (pv) and ferropericlase (fp). The observed anisotropy can be interpreted as due to lattice preferred orientation (LPO) of either pv, fp, or both in the lowermost mantle induced by vertical flow due to thermal buoyancy, which might be related to the origin of the Hawaiian hotspot (although other possibilities such as a chemically distinct layer, LPO of post-perovskite (ppv), or LPO in counter-flow in and around a chemically dense pile cannot be excluded). We show that resolution of the velocity of SV shear waves very close to the CMB is inherently limited due to the boundary condition of zero tangential traction at the CMB; shear wave splitting studies thus cannot be used to investigate the anisotropy of the lowermost mantle. In the presentation, we will also present and discuss results for other regions.
Recent progress in development and application of quantitative and objective inversion for body-w... more Recent progress in development and application of quantitative and objective inversion for body-wave waveforms has allowed us to obtain fine 1-D (vertical) structure for several regions (e.g., D" layer studies, Kawai et al., 2007ab, GRL). The inverse problem is written as: A δ m = δ d_OBS-INIT, [eqn:observation] where A is the matrix of partial derivatives, δ d_OBS-INIT is the residual of observed waveforms and synthetic waveforms for an initial model, and δ m is the perturbation to the initial model. We minimize the residual of the observed and the synthetics to obtain a modified model. Utilizing a dataset of many waveforms taken as a whole, we can extract robust information on the Earth's structure, even though there might be a relatively large difference between the observed data and the initial synthetics. The robustness of the inversion results is confirmed in several ways. In our inversion we use the SVD (singular value decomposition) for the matrix A: A = U Λ VT and VT V = [ δ m1 δ m2 δ m3 ·s ]. We represent the modified model δm as: δ m = M1δ m1 + M2δ m2 + ·s + Mnδ mn. Here we introduce a factor which we call "SVD sensitivity for each datum," which is the inner product of each observed waveform and the modified waveform for an individual model change. Plotting these factors helps us to visualize the distribution of the dataset and to extract robust information on regional structure. As we have mainly studied the mantle transition zone beneath Japan and environs, we find some structure which might represent the cold slab stagnating into the transition zone near Hokkaido, or lower Qμ for all the regions of our interest than the global average PREM model, which might be connected to thermal or compositional structure there.
We invert shear-wave waveform data for the radial variation of (isotropic) shear-velocity in D″ b... more We invert shear-wave waveform data for the radial variation of (isotropic) shear-velocity in D″ beneath Northern Asia. We reduce source and receiver effects by using data for intermediate and deep events beneath Italy and Japan recorded respectively at stations in East Asia and Europe. Relative to PREM, we find a significantly higher S-wave velocity in the depth range from 150 to 300 km above the core-mantle boundary (CMB) and a slightly lower S-wave velocity in the depth range 0-150 km above the CMB. As our previous studies of D″ structure beneath Central America and the Arctic obtained similar S-wave velocity models, we suggest that this pattern of vertical dependence of shear wave velocity in D″ may be a general phenomenon, at least in relatively cold regions.
Inversion of body-wave waveform data for elastic and anelastic transition-zone structure in and a... more Inversion of body-wave waveform data for elastic and anelastic transition-zone structure in and around Japan Nobuaki Fuji, Kenji Kawai and Robert J. Geller We have developed new methods to invert seismic waveform data for localized seismic structure. We use these methods to invert for the fine structure of the mantle transition zone in and around Japan using broadband data from regional arrays. Our methods use tools we have developed for performing 'static corrections' for the complex crustal structure in and around Japan. We conduct static correction by making a time shift which gives the best correlation coefficient for the first arrival S phase. We also have used other methods for time shifting such as onset pick time shift, but we found that the results of the inversion do not greatly differ. In order to stabilize the inversion we use automated data selection criteria, which limit the amplitude ratio of the synthetic seismogram to the observed seismogram to be between 0.3 and 3.0, and require a value above 0.5 for the correlation coefficient. The amplitude ratios (synthetic/observed) are systematically large for the initial (PREM) anelasticity model, so we invert simultaneously for both Q and elastic structure. We invert the transverse components of long-period (20-200 s) body-wave data from the NIED F-net array. The target regions lie beneath Hokkaido and the Philippines Sea at depths between 200 km and 700 km. The dataset used in this study consists mainly of triplication S phases that sample the mantle transition zone. It is difficult to analyze such data using previous methods, but our methods can handle such data. Models from studies of this type should contribute to improving our knowledge of the thermal state and proportion of water present in the upper mantle.
Using localised body-wave waveform inversion method for both elastic and anelastic structure (whi... more Using localised body-wave waveform inversion method for both elastic and anelastic structure (which we have recently developed), we found relatively high attenuation in the mantle transition zone beneath in and around Japan. We divide the region of interest into several sub-regions and obtained 1-D structure model both for elastic and for anelastic structure. From plural aspects of statistics (stacked waveforms, AIC, data distribution for kernels), we confirm that the high attenuation is the key to realise the amplitudes of the waveforms particularly (Fuji et al., in prep.). In this presentation, we will mention on our methodology briefly and show our results. Beneath the mantle wedge (Philippines Sea), we found higher attenuation than beneath the Pacific Ocean. This will suggest some anomaly information either on the temperature or on the water content. We will discuss the meanings further in the last of the presentation.
73rd EAGE Conference and Exhibition - Workshops 2011, 2011
In the last decade, the deployment of dense regional arrays such as the USArray transportable arr... more In the last decade, the deployment of dense regional arrays such as the USArray transportable array has considerably improved our capacity to image the interior of the Earth. However, the use of the wealth of information coming from these new and large amounts of high quality ...
We are carrying out small scale pilot efforts to invert actual data for the fine structure of D&q... more We are carrying out small scale pilot efforts to invert actual data for the fine structure of D" using waveform inversion. These efforts will then be significantly scaled up. Now, we consider broadband transverse component S-wave waveform data recorded in North America for deep events under South America. These events sample the lowermost several hundred km of the mantle (the D" region), immediately above the core- mantle boundary (CMB) under Central America. Interest in the D" region has recently become even more intense due to the discovery of a high-pressure phase (post-perovskite, abbreviated as "ppv"), which is likely to be responsible for the jump in seismic velocity at the top of D". For the moment our preliminary inversion seeks a laterally homogeneous isotropic S-velocity model appropriate to the base of the mantle under Central America. Since the inversion is only for the structure of D" in the target region, other effects must be account...
We present a method for calculating Fréchet kernels from a precomputed database of strain Green&#... more We present a method for calculating Fréchet kernels from a precomputed database of strain Green's tensors by the Direct Solution Method, which allows us to go to frequencies up to 1 Hz. In order to accelerate the kernel calculations, we interpolate and reconstruct the Green functions for arbitrary epicentral distances from the Green's tensor database. The algorithm is very efficient and allows us to calculate waveform, traveltime, and amplitude sensitivity kernels for high-frequency teleseismic body waves on a laptop. We will show travel-time and amplitude sensitivity kernels for high frequency P, PKP and Pdiff phases. We will also present waveform Fréchet derivatives in the distance range corresponding to the mantle transition zone discontinuities triplication and to PcP precursors. The temporal dependence of these Fréchet derivatives shows a high-degree of complexity, that can be exploited to resolve fine scale structures by waveform inversion.
In the last decade, the deployment of dense arrays such as the USArray transportable array has co... more In the last decade, the deployment of dense arrays such as the USArray transportable array has considerably improved our capacity to image the interior of the Earth. However, the exploitation of the wealth of information coming from these new and large amounts of high quality broadband data still heavily relies on asymptotic approaches. Here we present a new method that couples a spectral element method inside a regional 3-D domain with a global propagation approach in a spherically symmetric Earth model. This method is both very accurate and very efficient, even at periods as short as 1 s. It allows us to model short period teleseismic waves in 3-D media, and to compute 3-D waveform Fréchet derivatives, opening the way to 3-D waveform inversion of teleseismic body wave records from dense, local or regional, broadband arrays.
ABSTRACT Inversion of body-wave waveform data for elastic and anelastic transition-zone structure... more ABSTRACT Inversion of body-wave waveform data for elastic and anelastic transition-zone structure in and around Japan Nobuaki Fuji, Kenji Kawai and Robert J. Geller We have developed new methods to invert seismic waveform data for localized seismic structure. We use these methods to invert for the fine structure of the mantle transition zone in and around Japan using broadband data from regional arrays. Our methods use tools we have developed for performing 'static corrections' for the complex crustal structure in and around Japan. We conduct static correction by making a time shift which gives the best correlation coefficient for the first arrival S phase. We also have used other methods for time shifting such as onset pick time shift, but we found that the results of the inversion do not greatly differ. In order to stabilize the inversion we use automated data selection criteria, which limit the amplitude ratio of the synthetic seismogram to the observed seismogram to be between 0.3 and 3.0, and require a value above 0.5 for the correlation coefficient. The amplitude ratios (synthetic/observed) are systematically large for the initial (PREM) anelasticity model, so we invert simultaneously for both Q and elastic structure. We invert the transverse components of long-period (20-200 s) body-wave data from the NIED F-net array. The target regions lie beneath Hokkaido and the Philippines Sea at depths between 200 km and 700 km. The dataset used in this study consists mainly of triplication S phases that sample the mantle transition zone. It is difficult to analyze such data using previous methods, but our methods can handle such data. Models from studies of this type should contribute to improving our knowledge of the thermal state and proportion of water present in the upper mantle.
By using waveform inversion to study the seismic velocity structure of D'', we found th... more By using waveform inversion to study the seismic velocity structure of D'', we found that the upper half of D'' beneath Central America has S-wave velocity significantly in excess of PREM, while the lower half has S-velocity nearly equal to PREM [Kawai et al., GRL, 2007]. We interpret this as evidence for a double crossing phase transition (a reverse transition from post-perovskite=ppv to perovskite=pv), the existence of which was suggested by Hernlund et al. [2005, Nature]. In addition to the D'' layer beneath Central America, we have now also inverted seismic body- wave waveform data for the vertical dependence of (isotropic) shear-velocity in D'' beneath Asia and the Arctic, also using the transverse components of relatively long period broadband waveforms (20-200 s) as data. We found that these data also suggest the existence of high S-velocity in the upper half of D'' and low velocity in the lower half, consistent with the existence of a double crossing phase transition within D''. Our results for these three separate regions, when taken together, suggest that the existence of a double crossing phase transition is widespread and possibly global or nearly so. This has important implications for studies of the mineral physics, temperature profile, convection and material transport in D''. In the near future, we plan to study D'' beneath Africa and the Pacific in order to further determine the extent to which the double crossing phase transition is (or is not) ubiquitous.
ABSTRACT Recent progress in development and application of quantitative and objective inversion f... more ABSTRACT Recent progress in development and application of quantitative and objective inversion for body-wave waveforms has allowed us to obtain fine 1-D (vertical) structure for several regions (e.g., D" layer studies, Kawai et al., 2007ab, GRL). The inverse problem is written as: A delta m = delta d_OBS-INIT, [eqn:observation] where A is the matrix of partial derivatives, delta d_OBS-INIT is the residual of observed waveforms and synthetic waveforms for an initial model, and delta m is the perturbation to the initial model. We minimize the residual of the observed and the synthetics to obtain a modified model. Utilizing a dataset of many waveforms taken as a whole, we can extract robust information on the Earth's structure, even though there might be a relatively large difference between the observed data and the initial synthetics. The robustness of the inversion results is confirmed in several ways. In our inversion we use the SVD (singular value decomposition) for the matrix A: A = U Lambda VT and VT V = [ delta m1 delta m2 delta m3 ·s ]. We represent the modified model deltam as: delta m = M1delta m1 + M2delta m2 + ·s + Mndelta mn. Here we introduce a factor which we call "SVD sensitivity for each datum," which is the inner product of each observed waveform and the modified waveform for an individual model change. Plotting these factors helps us to visualize the distribution of the dataset and to extract robust information on regional structure. As we have mainly studied the mantle transition zone beneath Japan and environs, we find some structure which might represent the cold slab stagnating into the transition zone near Hokkaido, or lower Qmu for all the regions of our interest than the global average PREM model, which might be connected to thermal or compositional structure there.
ABSTRACT Using localised body-wave waveform inversion method for both elastic and anelastic struc... more ABSTRACT Using localised body-wave waveform inversion method for both elastic and anelastic structure (which we have recently developed), we found relatively high attenuation in the mantle transition zone beneath in and around Japan. We divide the region of interest into several sub-regions and obtained 1-D structure model both for elastic and for anelastic structure. From plural aspects of statistics (stacked waveforms, AIC, data distribution for kernels), we confirm that the high attenuation is the key to realise the amplitudes of the waveforms particularly (Fuji et al., in prep.). In this presentation, we will mention on our methodology briefly and show our results. Beneath the mantle wedge (Philippines Sea), we found higher attenuation than beneath the Pacific Ocean. This will suggest some anomaly information either on the temperature or on the water content. We will discuss the meanings further in the last of the presentation.
ABSTRACT The lowermost mantle (which is called the D'' region) is a boundary laye... more ABSTRACT The lowermost mantle (which is called the D'' region) is a boundary layer for mantle convection and is also a boundary layer between the outer core (composed primarily of molten iron) and the mantle, where compositional differentiation and thermal/chemical interaction can occur. Experiments by Murakami et al. (2004) found a phase transition from perovskite structure to post-perovskite structure at pressures corresponding roughly to the top of D''. Better knowledge of the seismic fine structure of D'' region thus can make an important contribution to understand the dynamics of the lowermost mantle. We have developed and implemented methods for inversion of seismic waveforms for Earth structure, and have applied them to data from dense seismic arrays (e.g., Kawai et al., 2007a,b, GRL). In a recent application of these methods, Konishi et al. (2009, EPSL) found evidence that basaltic crust is subducted to the lowermost mantle, on the basis of the fine S-wave velocity structure of D" beneath the Western Pacific. In this study, we conduct waveform inversion for the fine shear velocity structure in the lowermost mantle beneath Central Asia using data from ORFEUS and IRIS for earthquakes that occurred in 1995-2006 beneath Southeast Asia with epicentral distances of roughly over 75 degree. We use data bandpass filtered from 8-200s to invert for the vertical dependence of the shear velocity. We obtain a model whose the average shear velocity is faster than PREM in D''. We are conducting further inversions to obtain more information on this region using higher frequency data and more detailed regionalization.
ABSTRACT It is impractical to directly constrain the elastic constants of a transversely isotropi... more ABSTRACT It is impractical to directly constrain the elastic constants of a transversely isotropic (TI) medium using travel-time data. In contrast, as we show in this study, the elastic constants can be determined straightforwardly by inversion of body-wave waveform data. We invert the horizontal components of observed seismic waveforms of S and ScS phases (as well as any other phases arriving in the time window) to determine the radial profile of TI shear wave velocity in the lowermost mantle beneath the Pacific. We find that the radial (SV) component is faster than the transverse (SH) component in the depth range from about 200--400~km above the core-mantle boundary (CMB). The major mineralogical components above the D" discontinuity in this depth range are Mg-perovskite (pv) and ferropericlase (fp). The observed anisotropy can be interpreted as due to lattice preferred orientation (LPO) of either pv, fp, or both in the lowermost mantle induced by vertical flow due to thermal buoyancy, which might be related to the origin of the Hawaiian hotspot (although other possibilities such as a chemically distinct layer, LPO of post-perovskite (ppv), or LPO in counter-flow in and around a chemically dense pile cannot be excluded). We show that resolution of the velocity of SV shear waves very close to the CMB is inherently limited due to the boundary condition of zero tangential traction at the CMB; shear wave splitting studies thus cannot be used to investigate the anisotropy of the lowermost mantle. In the presentation, we will also present and discuss results for other regions.
Uploads
Papers by Nobuaki Fuji