S52B-01
Characterizing source properties of events in Southern Africa
Southern Africa has diffuse and relatively low levels of seismicity, except at the southernmost active part of the East African Rift zone in Mozambique, and around active mining areas in South Africa. Many of the mining induced events have depths of less than 4 kilometers, offer the potential for diverse focal mechanisms, and occur at a rate of tens of magnitude 3+ events per year. These characteristics make a good data set to study depth and mechanism effects on local and regional waveforms. We put together a dataset of 165 magnitude 3 and greater events from 1997-2006 from across Southern Africa recorded at the permanent stations BOSA,SUR and LBTB. We calibrated these stations for coda derived source spectra and moment magnitude using the regional envelope methodology of Mayeda et al. (2003). We extend the technique using a revised geometrical spreading path term (Phillips pers. Comm.). We tie the coda Mw to independent values from waveform modeling such as Harvard CMT and Bowers (1997). The resulting spectra are then examined to look at any evidence of systematic behavior with source depth, focal mechanism and any differences between tectonic events and the mining induced events. We are starting to perform regional waveform modeling of some of the larger mining events to provide additional moment values and to determine focal mechanisms. We are also using newly deployed seismic stations of AfricaArray around the mining areas to provide local ground truth information on mine event depth and mechanism.
S52B-02
An Improved Global Energy to Moment Ratio Catalog: More Events, Stations, and Rupture Durations
To enhance a global dataset of radiated earthquake energies, E, we use the methodology of Newman & Okal [1998] to develop an improved catalog of events that include activity for the past 10 years, incorporating all earthquakes with MW > 6.6. This new catalog contains 5 × the number of earthquakes, and 20 × the waveforms of the original study. For each event, E is compared to the seismic moment, M0, to determine average values and variations in earthquake strong shaking. In addition to a denser event database, we also extend the original dataset upward by a unit magnitude due to the inclusion of recent, very large earthquakes. Because we utilize the complete rupture durations (especially important for large and "tsunami earthquakes"), energy estimates more accurately represent true shaking for long duration events. We find that globally the average ratio, Θ = log10E/M0 = -4.66, is somewhat higher than the value (-4.98) found by Newman & Okal [1998]. Likewise, because of a vast improvement in globally available seismic data for more recent events, we can now better explore the effects of source directivity, together with random and path- dependent errors in global energy calculations. Additionally, we explore the possibility of regional effects on the E/M0 ratio, determining energies for all earthquakes MW>5.5 along the Middle America Subduction zone, the site of an energy-deficient "tsunami earthquake" in 1992. We find that regional dip-slip events are characterized by a lower average Θ of -5.13, with no significant along-strike observed changes. The lower regional value suggests that these events have generally slower ruptures, lower stress drops, or are in a higher attenuation environment than the global average.
S52B-03
Characteristics of Recent Moderate-Magnitude Seismic Activity in the Gulf of Mexico
On 10 September 2006, a reverse faulting earthquake occurred in the north central Gulf of Mexico (USGS Source parameters: Ms 5.5, mb 5.9, 14:56:08.16 UTC, 26.32N, -86.61E,14 km; GCMT Mw 5.8, 30 km depth). Although close to 300 km from the coast, the event was felt widely across the southeastern coastal region of the conterminous United States (particularly Florida) and parts of coastal Mexico and the Bahamas. Earlier that year, on 10 February, a slightly smaller moderate magnitude seismic event occurred beneath the continental slope south of Louisiana and Alabama (USGS Source parameters: Ms 5.2, mb 4.1, 27.597, -90.163, 5 km). This event was also felt in coastal regions. These events, the largest in "stable" region of the Gulf since at least 1973, did not occur on well known fault systems but in regions with at best sparse seismic activity. Since these events are not associated with well known structures, rare events of similar size may conceivably occur elsewhere in the Gulf region. Similar-sized events closer to the Gulf Coast economic infrastructure represent a potential seismic hazard from either shaking-induced damage or slope instability. We will present an analysis of the body-wave signals from these two events to compare and contrast their depths and source characteristics.
S52B-04
Comparison of Early Aftershocks for the 2004 Mid-Niigata and 2007 Noto Hanto Earthquakes in Japan
We compared the aftershock sequences of the similar 2004 Mid-Niigata (Mw6.6) and 2007 Noto Hanto (Mw6.7) earthquakes in central Japan. Although the two mainshocks had similar size, depth, and focal mechanisms, the numbers of aftershocks were quite different, with the Niigata mainshock producing a much stronger sequence. We examined the continuously recorded data from nearby Hi-Net stations operated by the National Institute for Earth Science and Disaster Prevention (NIED), to identify the early aftershocks following both mainshocks. A 5 hz high-pass filter was chosen to facilitate identification of the high-frequency arrivals from individual aftershocks. We used 6 stations distributed at distances within about 30 km. Aftershocks were identified by looking at large printouts of the continuous records for the six stations and peak amplitudes were measured to calculate the magnitude. The magnitude determination using these high-pass filtered records was calibrated by using a set of 30 earthquakes that were also listed in the catalog of the Japan Meteorological Agency (JMA). We estimate that the completeness level of small aftershocks is about Mj3.5. The event counts show that the aftershock sequences of the two earthquakes were quite similar for about the first 7 minutes. Following that time, the Niigata aftershocks clearly continue at a much higher rate which is about 3 times the rate of the Noto earthquake. The time where the rates diverge corresponds to the occurrence of a Mj6.3 earthquake in the Niigata sequence. This pattern can be seen in both the plots for the Mj¡Ý3.5 and Mj¡Ý4.0 events. Since there are more earthquakes for the Mj¡Ý3.5 data set, the time resolution is better. These results show an enhanced triggering of aftershocks for the Niigata sequence several minutes after the mainshock. The Niigata region is an area of hydrocarbon production with regions of high pressure fluids, and Sibson (2007) proposes that the swarm-like behavior is due to upward discharge of fluids from a deeper overpressured region. Fluid may flow into openings of fault zones caused by strong earthquake shaking, resulting in a reduction of the normal stresses on the faults.
S52B-05
Seismological Detection and Analysis of Recent Landslides in Alaska and the Yukon
Large landslides and avalanches generate seismic waves that can be used to detect, locate, and constrain the dynamic processes active in the slides. We have detected and located several landslide events using the Global Seismographic Network (GSN) as a long-period array. The detected events have equivalent long-period magnitudes of about M=5, while local magnitudes for these events, where available, are 2--3 magnitude units smaller. Two of the largest detected landslides are the 2005 Mount Steller (M=5.2) and the 2007 Mount Steele (M=5.2) events, located in the Pacific Coast Range in southern Alaska and the southern Yukon, respectively. In contrast to the forces active in standard earthquakes, in which seismic waves are generated as a consequence of tectonic stress drop within the Earth, landslides excite seismic waves through the time-varying forces caused by the acceleration and deceleration of a sliding mass interacting with the Earth's surface. Both the Mount Steller and the Mount Steele events involved sliding volumes of tens of millions of cubic meters of debris, vertical drops of around 2000 meters and runouts of more than 5 km. We use seismograms from several local and regional seismometers, including records from the STEEP PASSCAL array at epicentral distances as short as 10 km, to model the sliding process. We parameterize the source as a point force acting on the Earth's surface and obtain its magnitude, strike, and dip as a function of time. We observe the initial, nearly vertical, unloading force as the rock detaches from the solid Earth, a downward impulse as the mass is diverted into a horizontal trajectory at the base of the steep mountain slope, and the horizontal force corresponding to the deceleration phase as friction brings the sliding mass to a halt. Both slides have total durations of approximately 100 seconds. We find good agreement between the force histories inferred from seismograms and simple forward calculations of the dynamics of the sliding mass based on local topography, friction, and conservation of momentum.
S52B-06
Investigations of Icequake Source Physics - Assessing the Link Between Englacial Fracturing and Glacial Floods.
We study the seismic activity recorded on the Gornergletscher, Switzerland, during annual glacial flooding sequences. Historically, these 'jokulhaup' events have caused considerable damage downstream in the Zermatt Valley. As part of a wider monitoring effort of the glacier, seismic networks have been deployed each spring during 4 recent lake drainages of Gornersee, a glacier-dammed lake at the confluence of the glacier's two main tributaries. We observed several thousand seismic events within the glacier ice ('icequakes') each day. Most occur near the surface and are associated with crevassing. However, a small fraction of events can be reliably located at intermediate depths and near the glacier bed. In particular, we are interested in investigating the relationship between the sudden subglacial passage of large volumes of water and the observed deep icequakes. The temporal occurrence of these basal events indicates a relationship to the lake drainage, and we suggest they may be caused by large fluctuations in subglacial water pressure. However, information about source type, strength and fault orientation are critical to a more complete understanding of these deep icequakes. We estimate the moment tensor of these deep events using a full waveform inversion based on the method of Dreger (1994), with and without constraint on the isotropic component. We modify the inversion scheme, usually applied to regional seismicity, to the scale of the glacier and the high frequency of the observed signals by scaling distances and frequencies to match simple 1-D Green's functions derived at longer periods. The results of initial icequake inversions exhibit good waveforms fits: we match both deep event reflectivity synthetics as well as observed first motion focal mechanisms for clearly double-couple surface events. This shows that moment tensor estimations can be tailored to the glacial environment and can elucidate the icequake source physics. If the fracture processes of the deep icequakes can be associated with the flooding sequence both in terms of their temporal occurrence and their source mechanism, icequake monitoring can potentially be used to provide a key early warning to the impending flood.
S52B-07
Near Real Time Seismic Discrimination at Regional Distances Using a Voting Scheme of Discriminators Over the Lg Phase
The recent installation of telemetered BB stations as part of the National Seismic Network in Spain makes necessary to discriminate in the low magnitude range (around ML=2) between earthquakes and chemical explosions. The process should be in real time because an automated seismic bulletin is sent to different agencies. The discrimination processes uses a voting scheme based on the experience gained with a training set of 200 events which by a independent way is known its natural or artificial condition. The discriminators were applied to the Lg phase which is well developed at regional distances and were: spectral variance, spectral slope, relation variance-slope, and the logarithm of the ratio of amplitudes in the frequency bands, 1-2 hz/6-8 hz and 1-2 hz/7-9 hz respectively. The weights assigned to each discriminator are based in percentage of skill as compared with true classification. The most powerful index is the amplitude ratio 1-2 hz/ 7-9 hz and the lowest is the variance of the Lg phase. As new set of data is entered the system is trained as soon as the true classification is confirmed.
S52B-08
Resolving Isotropic Components from Regional Waves using Grid Search and Moment Tensor Inversion Methods
We applied the moment tensor (MT) analysis scheme to identify seismic sources using regional seismograms based on the representation theorem for the elastic wave displacement field. This method is applied to estimate the isotropic (ISO) and deviatoric MT components of earthquake, volcanic, and isotropic sources within the Basin and Range Province (BRP) and western US. The ISO components from Hoya, Bexar, Montello and Junction were compared to recently well recorded recent earthquakes near Little Skull Mountain, Scotty's Junction, Eureka Valley, and Fish Lake Valley within southern Nevada. We also examined "dilatational" sources near Mammoth Lakes Caldera and two mine collapses including the August 2007 event in Utah recorded by US Array. Using our formulation we have first implemented the full MT inversion method on long period filtered regional data. We also applied a grid-search technique to solve for the percent deviatoric and %ISO moments. By using the grid-search technique, high-frequency waveforms are used with calibrated velocity models. We modeled the ISO and deviatoric components (spall and tectonic release) as separate events delayed in time or offset in space. Calibrated velocity models helped the resolution of the ISO components and decrease the variance over the average, initial or background velocity models. The centroid location and time shifts are velocity model dependent. Models can be improved as was done in previously published work in which we used an iterative waveform inversion method with regional seismograms from four well recorded and constrained earthquakes. The resulting velocity models reduced the variance between predicted synthetics by about 50 to 80% for frequencies up to 0.5 Hz. Tests indicate that the individual path-specific models perform better at recovering the earthquake MT solutions even after using a sparser distribution of stations than the average or initial models.