ALBERT

Description: We have estimated P-wave and S-wave anelastic attenuation coefficients for the thick, unconsolidated sediments of the Mississippi embayment, central United States, using the spectral distance decay of explosion P and Rayleigh waves. The sediment-trapped P wave, P (sub sed) , is observed to ranges of 80 km at 10 Hz, and 1-Hz Rayleigh waves are observed out to 130 km from a 5000-lb borehole explosion in the northern part of the embayment. Rayleigh waves of 4 Hz are seen to distances of 3 km from a smaller 50-lb explosion. Analysis of the group velocity and amplitude-distance decay of both waves yields an average Q (sub s) of 100 and Q (sub p) of 200 for embayment sediments that are independent of frequency. Scatter in the Q estimates comes from interference of multiple P-wave reverberations and Rayleigh-wave modes. The attenuation model is self-consistent in that it is the same as obtained by the analysis of synthetic seismograms using the inferred Q-values. Inferred Q (sub p) and Q (sub s) values are more than three times higher than previous estimates and imply that unconsolidated sediments of the embayment do not significantly attenuate small-strain earthquake ground motions. These estimates represent a lower bound to Q of the sediments since significant scattering is observed in the waveform data that contributes to the distance decay of wave amplitude. Higher Q values also imply that the unconsolidated sediments of the embayment will form an efficient wave guide for surface waves radiated from shallow earthquakes or large earthquakes that rupture into the sediments, producing high-amplitude, long-duration wave trains that should be considered in earthquake hazard assessments.

Description: Seismic waveform data from the Cooperative New Madrid Seismic Network (CNMSN), the U.S. National Seismograph Network (USNSN), and stations operated by the U.S. Geological Survey and the Incorporated Research Institutions for Seismology (IU) are corrected to the theoretical Wood-Anderson response, and the peak amplitudes are used to determine a local-magnitude scale for the central United States. The distance-correction function can be expressed as -log A (sub 0) = 0.939 log (r/100)-0.000276 (r-100)+3.0, with amplitude A (sub 0) in millimeters and hypocentral distance r in kilometers, showing a weak attenuation with distance. Synthetic seismograms are calculated based on a detailed crustal model obtained from local well log acoustic data in the Mississippi embayment to examine the effect of structure on station factors. The results show that the Mississippi embayment sediment, as a whole, may amplify seismic waves at some specific frequencies, such as 0.5 and 5.0-6.0 Hz, with a factor less than 4.0, and may not amplify seismic waves in the frequency band from 0.5 to 9.9 Hz on average, even though the amplification factors do increase with the source magnitude. Estimations for b-values and magnitudes of earthquakes with a return period of 500 yr in the central United States are performed, based on the original New Madrid catalog (M (sub D) or m (sub bLg) ) and the newly created M (sub L) catalog. Results show that the b-value is smaller for the M (sub L) catalog (0.790) yielding one M 7.5 event every 500 yr, a result that converges with the paleoearthquake estimate for the region. On average, the relation between M (sub L) and M (sub D) or m (sub bLg) (for 1.5〈 or =M (sub D) 〈 or =5.0) can be expressed as M (sub L) =1.008M (sub D) +0.0714.

Description: The spatial displacement gradient of a seismic wave is related to displacement and velocity through two spatial coefficients for any one dimension. One coefficient gives the relative change of wave geometrical spreading with distance and the other gives the horizontal slowness and its change with distance. The essential feature of spatial gradient analysis is a time-domain relation between three seismograms that yields information on the amplitude and phase behavior of a seismic wave. Filter theory is used to find these coefficients for data from 2D areal arrays of seismometers, termed gradiometers. A finite-difference star is used to compute the displacement gradient for irregularly shaped gradiometers, and a relation for the frequency-dependent error in the displacement gradient is obtained and applied to ensure accurate estimates. This kind of array analysis is useful for gradiometers at any distance from a source and yields a variety of time-domain and frequency-domain views of wave-amplitude changes and horizontal phase velocity estimates across the gradiometer. For example, time-dependent horizontal slowness and wave-azimuth plots are natural results of the analysis. These time-domain maps may be used in conjunction with time-distance and horizontal slowness-distance models to locate seismic sources or may be used directly to study earth structure. These methods are demonstrated by using data from a small-aperture ( approximately 40 m) seismic gradiometer.

Description: A time-domain approach for solving for the change in geometrical spreading and horizontal wave slowness in wave gradiometry is presented based on the use of the analytic signal. The horizontal displacement gradient of a wave is linearly related to the displacement and its time derivative. The coefficients of this relationship give the change of geometrical spreading, the change in radiation pattern, and horizontal slowness. The new time-domain technique incorporates estimates of the instantaneous amplitude and frequency of the three time series to solve uniquely for the wave-field coefficients. The analysis is simpler and more suited to fast array processing of displacement gradient data sets compared with a spectral ratio method.

Description: A local magnitude (M (sub L) ) scale and seismicity catalog for the Ethiopian Plateau have been developed using data collected by the 2000-2002 Ethiopia Broadband Seismic Experiment. Locations for 253 local and regional events have been obtained using P-wave arrival times recorded on four or more stations. For constructing the M (sub L) scale, waveforms were corrected for instrument response and convolved with the nominal Wood-Anderson torsion seismograph response. Maximum S-wave amplitudes were then picked on the horizontal components of ground motion (3218 total observations) and used in an inversion for event magnitudes, two model parameters, and 54 horizontal component station corrections. The distance correction obtained from the inversion is -logA (sub 0) = 0.726 log(r/100)+0.000558(r-100)+3.0, where r is the hypocentral distance in kilometers. Seven of the 253 events can be considered ground truth (GT) events, with epicentral locations accurate to within 5 km according to the GT5 local criteria of Bondar et al. (2004). In contrast to previously reported ground-motion attenuation for the Main Ethiopian Rift, we find relatively low ground-motion attenuation for the Ethiopian Plateau, reflecting variations in crustal structure between the Ethiopian Plateau and Main Ethiopian Rift.

Description: Ambient ground-motion data were collected using phased seismic arrays in fall 2002 and spring 2007 within the Mississippi embayment and at a single station external to the embayment. These data allowed us to determine the wave-field composition of ambient noise for understanding wave-propagation mechanisms giving rise to spectral peaks using Nakamura's H/V technique. Ambient ground motions in the frequency band of 0.1-0.33 Hz (10-3 sec period) were dominated by spatially localized Rayleigh- and Love-wave microseisms generated by high-ocean waves along the North American seaboard in the time periods of analysis. Seismic waves important in forming the H/V peak near 4 sec period are composed of relatively high-phase velocity Rayleigh and Love waves that convert to propagating homogeneous shear waves in the thick unconsolidated sediments of the embayment. The H/V resonant period is controlled by both constructive and destructive interference of these shear waves. A simple relationship for the H/V peak is given using a propagator matrix formulation that predicts the resonance frequency of a layered medium for surface wave motion at the base of the system. The amplitude of the observed H/V peak, however, does not give an accurate estimate of shear-wave amplification because it depends on the slowness of the incident wave. The inconsistency in estimated average shear-wave velocities using the H/V method and differential travel times of local earthquake Sp phases in the Mississippi embayment may be explained by misidentification of Sp-wave conversion points from deeper interfaces.

Description: We utilize the characteristic features of primary P- and SH-wave coda and Sp waveforms from local microearthquake data and perform prestacking 1D migration in an attempt to image prominent reflectors in the upper crust of the New Madrid seismic zone (NMSZ). This methodology applies to data from broadband stations, PARM and PENM, of the Cooperative New Madrid Seismic Network (CNMSN). Nearby exploration well log acoustic data of the Wilson 2-14 and Dow Chemical/Wilson number 1 wells are used to constrain the upper 4 km of P-wave velocity model. Despite polarity differences among P, SH, and Sp waveforms, this technique demonstrates that consistent reflectors in the deep sedimentary section can be imaged commonly among the three wave types. There are excellent correlations associated with the base of the upper Cretaceous/Holocene Mississippi Embayment Supergroup and the base of Knox group, which were also reported in a study of nearby seismic-reflection profiles by Hamilton and Zoback (1982). The Bonne Terre formation seems to be a prominent seismic stratigraphic marker associated with an interface displayed in the profiling. We find that earthquake event S-P times must be approximately 3 sec or more to resolve reflectors at about 4 km depth. Possible basement at about 4.0 to 4.5 km appears on the reflected P-wave image for PARM and SH-wave image for PENM. We conclude that the velocity structure of the upper 4 km crust beneath PARM can be represented by a constant velocity of V (sub p) =2.165 km/sec and V (sub S) =0.70 km/sec for the unconsolidated sediments and V (sub p) =6.0 km/sec and V (sub S) =3.2 km/sec for the Paleozoic sedimentary rocks. A constant velocity of V (sub p) =2.15 km/sec and V (sub S) =0.615 km/sec for the unconsolidated sediments and V (sub p) =6.0 km/sec and V (sub S) =3.4 km/sec for the Paleozoic sedimentary rocks can represent structure beneath PENM.