S34A-01 INVITED
A future for drifting seismic networks
One-dimensional, radial Earth models are sufficiently well constrained to accurately locate earthquakes and calculate the paths followed by seismic rays. The differences between observations and theoretical predictions of seismograms in such Earth models can be used to reconstruct the three-dimensional wave speed distribution in the regions sampled by the seismic waves, by the technique of seismic tomography. Caused by thermal, compositional, and textural variations, wave speed anomalies remain the premier data source to fully understand the structure and evolution of our planet, from the scale of mantle convection and the mechanisms of heat transfer from core to surface to the international between the deep Earth and surface processes such as plate motion and crustal deformation. Unequal geographical data coverage continues to fundamentally limit the quality of tomographic reconstructions of seismic wave speeds in the interior of the Earth. Only at great cost can geophysicists overcome the difficulties of placing seismographs on the two thirds of the Earth's surface that is covered by oceans. The lack of spatial data coverage strongly hampers the determination of the structure of the Earth in the uncovered regions: all 3-D Earth models are marked by blank spots in areas, distributed throughout the Earth, where little or no information can be obtained. As a possible solution to gaining equal geographic data coverage, we have developed MERMAID, a prototype mobile receiver that could provide an easy, cost-effective way to collect seismic data in the ocean. It is a modification of the robotic floating instruments designed and used by oceanographers. Like them, MERMAID spends its life at depth but is capable of surfacing using a pump and bladder. We have equipped it with a hydrophone to record water pressure variations induced by compressional (P) waves. Untethered and passively drifting, such a floating seismometer will surface upon detection of a "useful" seismic event (for seismic tomography, that is), determine a GPS location, and transmit the waveforms to a satellite. In this presentation we discuss the progress made in this field by our group. More specifically, we discuss the results of preliminary tests conducted off-shore La Jolla in 2003 and 2004, as well as just-in results from a third successful, in situ, test completed in August 2007. We will draw attention to design issues and bottlenecks and the need for and features of sophisticated onboard data analysis software which we have developed and tested. We will chart a road map of the way to our ultimate goal: a worldwide array of MERMAID floating hydrophones, on the scale of the current international land-based seismic arrays. This, we believe, has the potential to progressively eliminate the discrepancies in spatial coverage that currently result in seismic Earth models that are very poorly resolved in places. http://www.frederik.net
S34A-02
Response of Seismometer with Symmetric Triaxial Sensor Configuration to Complex Ground Motion
Most instruments used in seismological practice to record ground motion in all directions use three sensors oriented toward North, East and upward. In this standard configuration horizontal and vertical sensors differ in their construction because of gravity acceleration always applied to a vertical sensor. An alternative way of symmetric sensor configuration was first introduced by Galperin (1955) for petroleum exploration. In this arrangement three identical sensors are also positioned orthogonally to each other but are tilted at the same angle of 54.7 degrees to the vertical axis (triaxial system of coordinate balanced on its corner). Records obtained using symmetric configuration must be rotated into an earth referenced X, Y, Z coordinate system. A number of recent seismological instruments (e.g., broadband seismometers Streckeisen STS-2, Trillium of Nanometrics and Cronos of Kinemetrics) are using symmetric sensor configuration. In most of seismological studies it is assumed that rotational (rocking and torsion) components of earthquake ground motion are small enough to be neglected. However, recently examples were shown when rotational components are significant relative to translational components of motions. Response of pendulums installed in standard configuration (vertical and two horizontals) to complex input motion that includes rotations has been studied in a number of publications. We consider the response of pendulums in a symmetric sensor configuration to complex input motions including rotations, and the resultant triaxial system response. Possible implications of using symmetric sensor configuration in strong motion studies are discussed. Considering benefits of equal design of all three sensors in symmetric configuration, and as a result potentially lower cost of the three-component accelerograph, it may be useful for strong motion measurements not requiring high resolution post signal processing. The disadvantage of this configuration is that if one of the sensors is not working properly or there is a misalignment of sensors, it results in degradation of all three components. Symmetric sensor configuration requires identical processing of each channel putting a number of limitations on further processing of strong motion records.
S34A-03
Unexpected Mode Observations in the Low Frequency Seismic Spectrum
The initial observations of the continuous excitation of Earth's normal modes were first presented a decade ago. While significant research by multiple groups has been directed at the source of the continuous excitation since these discovery papers, one set of interesting observations in the discovery papers has not been pursued. These observations are of modes or lines in the seismic spectrum which do not correspond to Earth normal modes. We present more detailed multitaper power spectral density observations of these modes in seismic data recorded by the IRIS Global Seismic Network during seismically quiet times. In addition to the generally accepted continuously forced Earth normal modes, we find the following observations: 1) Modes in seismic spectrum not associated with Earth normal modes. 2) Q (quality factor) of these observed modes have values 1000 and higher. 3) Amplitudes of these modes observed on horizontal seismometers is significantly larger than observed on vertical seismometers. 4) The frequencies and Q of these modes in the seismic spectrum are consistent with Solar normal modes We suggest that the ultimate source of these modes observed in the seismic spectrum are solar modes coupling into the terrestrial system.
S34A-04
Characteristics and Source of Unexpected Modes Observed in the Low Frequency Seismic Spectrum
The initial observations of the continuous excitation of Earth's normal modes were first presented a decade ago. While significant research by multiple groups has been directed at the source of the continuous excitation since these discovery papers, the proposed sources do not explain the observations. We present more detailed multitaper analysis of seismic data recorded by the IRIS Global Seismic Network, laser strainmeter data from PFO, and coherence estimated between this data and the interplanetary magnetic field measured at ACE. Our analysis shows: 1. Coherence and phase relations between the North and East components of seismic velocity are largely aligned with the geomagnetic field. This also occurs in data from laser strainmeters, excluding simple magnetic coupling to the instruments. 2. The temporal dependence of the hum exhibits a strong diurnal dependence aligned with geomagnetic activity. The lunar component is usually weaker or absent, so forcing by nonlinear interactions of ocean waves seems improbable. 3. The horizontal seismic velocity is coherent with the interplanetary magnetic field measured by the ACE spacecraft at the first Lagrange point and with magnetic fields measured on GOES and similar sources. We conclude that "Earth's hum comes from the Sun".
S34A-05
Excitation of Long Period Rayleigh Waves by Large Storms in the North Atlantic Ocean
Ten years after the discovery of the continuous excitation of the Earth's free oscillations (Nawa et al. 1998), often referred to as the hum of the Earth, the sources of this excitation are still under debate. Recent studies suppose that the Earth's hum is caused either by atmospheric turbulence or by infragravity waves in the oceans. While under normal conditions, it is not possible to identify 'hum events', Rhie and Romanowicz (2006) showed that large storms over the Northern Pacific ocean with high waveheights are able to produce long period (T>100 sec) Rayleigh waves when they reach the west coast of North America. These Rayleigh waves can be detected using broadband seismic networks as arrays. We recently demonstrated the suitability of the German Regional Seismic Network (GRSN) for array studies of the long period seismic background (Kurrle and Widmer-Schnidrig 2006). We focus on these data to provide further evidence for the capability of ocean storms to cause seismic waves at periods T>100 sec. Several storms in the Atlantic Ocean are examined, and data from the GRSN and the Berkeley Digital Seismic Network (BDSN) are used to infer the source regions of the seismic waves. Comparison of the seismic results with global waveheight distributions from Wave Watch 3 nowcasts, provided by the Fleet Numerical Meteorology and Oceanography Center (FNMOC), reveal the direct connection between the storms and the seismic waves.
S34A-06
Validation of a broadband source modeling method of interplate earthquakes
A method to model broadband source process of interplate earthquakes for ground motion prediction is examined through comparing predicted ground motion with observation for the 2003 Tokachi-oki, Japan, earthquake. Recent studies (e.g., Nagai et al. (2001)) show that asperities (patches with relatively large slip on a source fault) of interplate earthquakes are fixed over recurrent earthquakes. Following this idea, I assume that the long- wavelength feature of the slip distribution , i.e., asperity distribution, of a future earthquake is to be the same as what was derived by source inversion analysis. The shorter-wavelength feature, which is not constrained by the source inversion is added taking consideration of stochastic characteristics of earthquake source. Multi-scale heterogeneity is introduced to both the slip distribution and the rupture velocity distribution of the inverted source model by increasing or decreasing the values in randomly-distributed, various-size patches. The amount of the increase/decrease of the slip is set proportional to the radius of the patch and that for the rupture velocity is set to constant which gives adequate variation seen in many source inversion results. Finally, the amplitude of slip spectrum is adjusted to decay as k-1.75, which was derived by Mai and Beroza (2002) that conducted stochastic analysis of earthquake slip complexity using many published source models. A slip velocity time function approximated from dynamic rupture simulation (Nakamura and Miyatake, 2000) is assumed. The 2003, Tokachi-oki, Japan, earthquake occurred along the Kuril trench provided plenty of ground motion waveform data and the source process models are estimated using the waveform data. I construct a broadband source process model from an inverted source model by Aoi et al. (2006) and compare synthetics with the observation at sites less subjected to subsurface structure. The adequacy of deciding microscopic source parameters responsible to generation of high frequency components is examined.
S34A-07
Imaging the Western Pacific Subduction Zones Using High-resolution Radon Transforms
Underside reflections (SS, PP precursors) have been widely used in imaging seismic discontinuity structures since the early 1990s. While these waves provide a outstanding global coverage, some regions remain under- sampled, for example, much of the Western Pacific subduction zones and South America. Furthermore, the so called "cap centers" of common signal enhancement schemes do not always reflect the true centers of data, nor do they aim to reduce the effect of uneven coverage. We present results from analyzing the largest SS precursor dataset to date (from 1981-2007) along the Western Pacific subduction zones. This data set more than doubles the data used in the global study of Gu et al. (2003) in the same regions. We adopt variable cap sizes (hence variable resolutions) to equalize the data count, and apply a high-resolution Radon Transform method to simultaneously constrain the ray angle and timing of SS precursors. Our study shows strong depressions of the 660-km discontinuity along the subduction zones, and the lateral variations strongly correlate with regional seismic velocities. The lateral extent of the depressions is smaller than previously found, which we attribute to the high resolution of our new approach. The 410-km discontinuity is elevated beneath the northwestern Pacific, and the anti-correlation with the topography of the 660- km discontinuity reflect significant thermal variations. Overall, however, the correlation between the two transition zone discontinuities is poor in along the Western Pacific subduction zones. The presence of water, garnet phase transformation and the limited lateral dimension of subducted slabs could all affect the topography of these discontinuities, particularly the 410 km discontinuity. The thickness of the transition zone, however, appear to be a more robust measure of temperature: it is consistently narrower than the global average across these subduction zones, as would be expected from the olivine phase transitions. In addition to the transition zone discontinuities, our high-resolution Radon images also present clear evidence of mantle reflectors near 250 km and 850-950 km. The occurrence of the latter reflector is more common than previously proposed and could represent a potential global discontinuity. Phase transformation or chemical differentiation could contribute to this well-resolved interface.
S34A-08
Mapping southern Californian crust with high and low frequency seismics: A comparison between tomography, receiver functions, and reflection imageries
While it is not surprising that the Earth may yield different geophysical imageries at different frequency scales, how much similarity or difference is there between crustal and mantle heterogeneities of different spatial scales? What are the geologic and geodynamic implications? Of course, we have to weed out insufficiencies in data S/N ratio, data coverage, and processing methods. This study intends to provide a careful comparison between seismic imaging methods of different spatial scales in southern California where data is of the state-or-the-art quality due to the dense distribution of earthquakes and seismologic stations, as well as the use of both passive earthquake data and active Los Angeles Regional Seismic Experiment (LARSE) data. The methods include traveltime tomography of long-wavelength velocity variations, receiver functions of long-wavelength seismic discontinuities, and reflection imaging of seismic impedance contrasts or scatters using envelop stacking in comparison with prestack depth migration. The comparison indicates a broad level of similarity between seismic imageries of different wavelengths, while well-known and somewhat less-known artifacts are still a key factor to reckon with. Compared to previously published analyses of LARSE data, the prestack depth migration method can better treat lateral velocity variations in the shallow crust. Prestack depth migration also improves the resolution of reflection events in comparison with the receiver functions. Prestack depth imaging is superior than simple CMP stack techniques to process 2D seismic data in the presence of high noise level, strong lateral velocity heterogeneity and crooked survey geometry. The long-wavelength undulation of velocity contours as provided by a new deformable-layer tomography is very helpful to interpreting the reflection images. The innovative seismic imaging techniques from this study are directly applicable to EarthScope/USArray and other projects studying the structure of the crust and mantle.