S13B-1298
The 2003 M=6.9 Zemmouri, Algeria, Earthquake Brought Thrust and Strike-Slip Faults Near Algiers Closer to Coulomb Failure
We investigate key features of thrust earthquake triggering, inhibition, and clustering associated with the stress transferred by the 2003 M=6.9 Zemmouri quake on an offshore hidden thrust fault in coastal Algeria. A crucial question is whether the seismic hazard increased on the Boumerdes and Thenia faults, which lie just west of the Zemmouri rupture and only 10-20 km from the city of Algiers. The capital city suffered large damaging quakes in A.D. 1365 and 1716, and is today home to 3 million people. Slip on blind thrust faults tend to increase the stress above the source fault and in much of the surrounding crust, whereas slip on surface-cutting thrust faults drops the stress in most of the adjacent crust. We examined the sensitivity of the imparted stress to different published source models of the 2003 Zemmouri event inferred from geodetic and seismic inversions, and focus here on the robust results. We calculate that the 2003 M=6.9 Zemmouri quake brought the Coulomb stress 1.0 bars closer to failure on the reverse Boumerdes and 0.5 bars closer on the right-lateral Thenia faults that bound the populated Mitidja basin, although the Thenia fault may not be tectonically active. The calculated pattern of the stress increase appears consistent with aftershock distribution determined from double difference earthquake tomography by Ayadi et al. (submitted); both of these faults were illuminated by aftershocks during the first three months of the sequence. The East Sahel and Larbaa faults, which lie further to the west, are calculated to have sustained a weak 0.1-bar stress increase and show no associated aftershocks. We also calculate a 1.0-bar stress increase on the NNW-SSE trending vertical right-lateral Kabyle fault located south of the Zemmouri fault, although there is no evidence of recent Quaternary tectonic movement, no geomorphology typical of active zones, and little seismicity along the Kabyle fault.
S13B-1299
Static stress changes along strike slip faults in Slovenia and Greece and seismic hazard implications
We examine the Coulomb (static) stress pattern following moderate magnitude earthquakes along strike-slip faults in NW Slovenia during 1998 and 2004 and several strong earthquakes in the North Aegean Sea (Greece). The objective is to investigate the seismic hazard implications for these areas given that future earthquakes may be triggered as a result of stress changes along neighbouring faults. Our findings include: a) stress levels have increased along the active Ravne fault (Slovenia) for all models discussed b) stress levels have decreased along the active, NW-SE striking Idria fault and c) stress levels throughout the crust have increased along the E-W direction but have decreased in the N-S direction (stress shadow effect). We also observe a pronounced stress shadow effect in the central Aegean Sea following two large earthquakes during December 1981 that terminated in July 2001 with a large (Mw=6.4) event near Skyros island. We also mapped a better correlation of the off-fault aftershock locations with stress maps incorporating the regional stress field.
S13B-1300
Static Stress Transfer between the Chinshan and Sanchiao Faults in the Taipei Metropolitan Area
In the Taipei Metropolitan Area, there is a significant problem to be resolved: Can two major active faults, i.e., the Chinshan and Sanchiao faults, in the area fail simultaneously? The former is a thrust fault and the latter a normal one. Hence, static stress transfer between a normal fault and a thrust one is a significant problem. In this work, we study this problem by computing the static Coulomb failure stress changes on a fault plane due to a failure on the other. Numerical computations are conducted for the combinations of three different dip and rake angles of the two fault planes. Results show that the no matter which one break first, the other can be triggered in some conditions, and rupture propagates northward for the Chinshan fault and southward for the Sanchiao fault. A failure on a less steep normal (Sanchiao) fault is more capable of trigging a steeper thrust (Chinshan) fault, and vice verse.
S13B-1301
2007 Mw=6.6 Niigata Chuetsu-Oki earthquake ruptured on a fault strongly unclamped by the 2004 Mw=6.6 Niigata Chuetsu shock
The highly damaging 2004 and 2007 earthquakes struck 40 km apart in an active fold and thrust belt along the Japan Sea coast of Honshu. Both earthquakes slipped immature reverse faults that are probably reactivated normal faults. The 2004 event ruptured several antithetical faults, and produced an unusually vigorous and widespread aftershock sequence, which including four M≥6 events within the first week. We argue that the high rate of aftershocks results from slip on blind high-angle reverse faults. Unlike reverse faults that cuts the earth's surface, or low-angle thrust faults, slip on steeply-dipping blind reverse faults impart high stresses to the surrounding crust. This hypothesis raises the question as to whether the 2007 earthquake is a delayed aftershock or in some sense an event promoted by the 2004 earthquake. If the fault friction is high (coefficient ~0.8), we find that the 2007 fault was subjected to a Coulomb stress increase of up to 0.3 bars by the 2004 earthquake. However, the Coulomb stress increase is found only if the 2007 rupture occurred on the NW-dipping nodal plane (strike/dip/rake of 228°/42°/95°), and only if the hypocenter lies at depth greater than about 15 km. These conditions are reasonable, but not certain. Regardless of the depth and nodal plane, the 2007 fault was subjected to an unclamping stress, and this could be as large as 0.9 bars. Unclamping may be very important for faults with little cumulative slip, because they presumably have rough surfaces. It is also possible that crustal fluids were driven into the unclamped regions following the 2004 earthquake, slowly lubricating the 2007 and surrounding faults. Although much closer together, the 1983 Mw=6.7 Coalinga, California thrust earthquake strongly unclamped the nearby Nunez reverse fault, which ruptured in a Mw=6.0 earthquake 2 months later. Thus the strongest link between the two Niigata earthquakes is coseismic unclamping.
S13B-1302
Jet Stream Converges Prior to 6.8M Niigata Chuetsu-oki Earthquake of Japan on 2007/07/16
The 6.8M Niigata Chuetsu-oki earthquake occurred on 2007/07/16 and resulted in 11 deaths and at least 1000 injuries have been reported, and 342 buildings were completely destroyed. The 108km/hr isobar jet stream line converged around an epicenter on 2007/07/01 12:00 and 2007/07/02 06:00. Before a devastating earthquake occurs, the underground water level usually changes caused by the rock squishing or loosening. This study assumed that rock squishing or loosening caused air inhalation or exhalation that creates an internal gravity wave. This phenomenon will change the jet streams at an altitude of 10 km. Ps. The predicted Data:07/06/26-07/07/26 Japan(37.4N140.0E)M 6.0 100% The Actual Data: 07/07/16 Japan (37.576N138.469E) 6.6M 10km This earthquake prediction had been predicted on http://tw.myblog.yahoo.com/wu10002002/ and sent to Dr. Dimitar Ouzounov in advance.
S13B-1303
Search of Possible Triggered Seismicity Patterns of Northern Tien Shan
Statistical analysis of the Northern Tien Shan seismicity was performed considering the possible triggering impacts of natural and man-made mechanical and electromagnetic factors on seismic activity. Strong distant earthquakes, lunar-solar tides, and magnetic storms are considered as natural triggering factors. The man-made factors include underground nuclear explosions (UNE) and electromagnetic impacts provided by high-power magnetohydrodynamic pulsed (MHD) generators. The representative local earthquake catalog of the region under study (41°-46° N, 74°-82° E) includes 15577 events of M>1.67 from 1975 to 2000. Within this time period 330 UNE and 109 firing runs of MHD generators, which are considered as the possible man-made earthquake triggering factors, have been performed within or adjacent to the analyzed region. Various statistical methods (cross-correlation, spectral analysis, RTL-analysis, etc.) were employed. For the used statement of problem and applied initial data the statistically significant patterns of triggered seismicity of the Northern Tien-Shan due to impacts of UNE and MHD generators were not found. Large common periods of seismicity variation for time series of distant strong earthquakes and local seismic events were selected. There is significant number of common periods (7, 9, 14, 28, 186, and 16384 days) for variation of z-component of the earth tide and release of seismic energy that may point to an influence of the earth tides on the local seismicity.
S13B-1304
Susceptibility to Long-Range Earthquake Triggering in California and Japan
Earthquake triggering associated with the passing seismic waves of large distant earthquakes has been observed regularly and robustly in California, particularly in volcanic and hydrothermal areas. Japan is known for its Holocene volcanic and hydrothermal activity, yet little long-range triggering has been observed. In order to develop a more detailed understanding of the conditions required for long-range triggering, we map long-range triggering susceptibility in California and Japan and compare it to maps of background activity, aftershock productivity, and surface heat flow. A spatial correlation is found between all four quantities in California, but no consistent pattern is found in Japan. Long-range triggering is tentatively identified in Kyushu, and possibly western Hokkaido. We find that Japan is less susceptible than California to short-range triggering, in that shallow crustal (<15km depth) earthquakes in Japan have on average fewer aftershocks than equivalent earthquakes in California. We infer that aftershock productivity and long-range triggerability are controlled by local fault properties in California and that tectonic style must influence triggering susceptibility. Triggerable regions in both California and Japan are predominantly trans-tensional and hydrothermal. Long-range triggering maps were produced by collecting all local events in the days before and after potential global triggering events, here defined as events that exceed an arbitrary magnitude-distance threshold and are located beyond some minimum distance. Raising the threshold restricts the analysis to stronger triggers but reduces sample size; we attempt to strike a qualitative balance between signal strength and robustness. We Regional triggering susceptibility is mapped by binning and differencing pre- and post-trigger seismicity and applying a gaussian smoothing filter. Short-range triggering susceptibility (traditional aftershock productivity) is mapped by counting the average number of early aftershocks generated by isolated mainshocks. We designate earthquakes as mainshocks if they are preceded by no larger events within 250km in the last 24 hours, and identify earthquakes as aftershocks if they fall within 15km and 1 hour of a mainshock. We bin and average the productivities of the mainshocks using the same smoothing scheme as used for the long-range map. The ANSS global catalog since 1984 was used for mapping in California and the JMA catalog since 1997 was used for Japan. Heat flow data is from the Southern Methodist University database and the Geological Survey of Japan. The maps show that long-range triggering thresholds are controlled by local fault conditions and that these conditions also influence traditional short-range aftershock productivity. Hydrothermal activity and trans-tensional tectonics are implicated as important conditions for triggering.
S13B-1305
Coseismic Stress Changes of the September 28, 2004 M=6 Parkfield Earthquake From Variation of Recurrence Time of Microearthquakes
We relocated a set of earthquakes that occurred between January 1984 and May 2007 in the vicinity of Parkfield, California. Space-time correlations performed on earthquakes that occurred before the September 28, 2004, M=6 earthquake reveal : (1) a triggering mechanism that dominates the dynamics over a period of about 10 days, (2) a cut-off in triggering at < 1 rupture length but with no clear-cut shadow and (3) an enhanced triggering at 100- 1000 days (i.e. multiplet recurrence times). The unique framework of the Parkfield seismicity (precise earthquake relocations, repeaters) allows to estimate stress changes at a very small scale. Computation of the clock advance caused by the M=6 mainshock for individual multiplet leads to an estimation of the stress changes at the scale of the multiplet size. This reveals insight of the stress heterogeneities on the fault plane caused by the 2004 mainshock.
S13B-1306
Dynamic triggering of high-frequency bursts by strong motions during the 2004 Parkfield earthquake sequence
High-pass filtering (30Hz) of acceleration records from the USGS Parkfield Dense Seismograph Array (UPSAR) reveals a series of bursts that occur only during the strong shaking of the 2004 M6 Parkfield earthquake (Mw 6.0) and its large aftershocks. These high-frequency bursts are probably associated with events occurring very near each station, and are dynamically triggered by the strong ground motions of large earthquakes. Supporting evidence includes a lack of correlation for bursts at closely spaced stations, and a sharp decrease of array coherency with increasing frequency and inter-station distance. Lack of correlation may be associated with the extremely high attenuation in the highly fractured layer that compromises the top few hundred meters of crust. The threshold stress required for triggering was found to be 0.16 to 6.5 MPa with a high signal-to-noise ratio, consistent with a previous estimate based on strong motion data from the 1999 Chi-Chi earthquake in Taiwan [Fischer et al., 2007, in review].
S13B-1307
Investigating the Effects of Stress Interaction Using a Cellular-automaton Based Model in Fault Networks of Varying Complexity.
Seismicity spatial and temporal patterns are strongly influenced by stress interaction between faults. However the effects of such interaction on earthquake statistics is not yet well understood. Computer models provide accurate, large and complete datasets to investigate this issue and also have the benefit of allowing direct comparison of seismicity behavior in time and space in networks, with and without fault interaction. We investigate the effect of such interaction on modeled real-world fault networks of varying complexity using a cellular-automata based model. Each 3-D fault within the fault network is modeled by a discrete cellular automaton. The cell size is 1 km square which allows for a minimum earthquake size of approximately Mw=4. The cell strength is distributed fractally across each fault and all cells are loaded by a remote tectonic stressing rate. When the stress on a cell exceeds its strength, the cell fails and stress is transferred to its nearest neighbors which may in turn cause them to break allowing the earthquake to grow. These stress transfer rules allow realistic stress concentrations to develop at the boundary of the rupture. If the extent of the rupture exceeds a user defined minimum length, and if interaction between faults is allowed, a boundary element method is used to calculate stress transfer to neighboring faults. Here we present results from four simulated fault networks based on active faults in the San Francisco Bay Area, California, the Northern Anatolian Fault, Turkey, Southern California, and the Marlborough Fault System, South Island, New Zealand. These are chosen for their varying level of fault complexity and we examine both interacting and non-interacting models in terms of their b-value and recurrence intervals for each region. Results will be compared and discussed.
S13B-1308
Rate-State Modeling of Stress Relaxation in Geometrically Complex Fault Systems
Slip of geometrically complex faults involves interactions and processes that do not occur in standard planar fault models. These include off-fault yielding and stress relaxation, which are required to prevent the development of pathological stress conditions on the fault (or in extreme cases fault lock-up). Nielsen and Knopoff [1988] introduced yielding through a simplified form of viscoelastic stress relaxation. However, the mechanical characteristics of the brittle seismogenic crust indicate that faulting processes will dominate the stress relaxation processes. The fractal-like character of fault systems and fault roughness, together with the finite strength of rocks, insures that slight movements of secondary faults, at all scales, will be necessary to accommodate slip of major through-going faults. To model the integrated effect of these processes, we employ an earthquake rate formulation [Dieterich, 1994], which incorporates laboratory-derived rate- and state-dependent frictional properties, on geometrically complex faults. With the rate-state formulation we find that stress relaxation occurs co-seismically during large earthquakes, as delayed stress relaxation in the form of aftershocks, and as spatially distributed background seismicity. During aftershocks the spatial mean of stresses decay at a rate proportional to 1/t. We find large spatial and temporal differences in models of slip of faults with relaxation compared to faults in purely elastic media. We conclude that that yielding and relaxation are important controlling processes that are the mechanics of slip on geometically complex faults
S13B-1309
Effect of 3D Stress Heterogeneity on Aftershock Sequences
Observations of spatially varying slip along fault zones and in earthquakes suggest that both slip and stress are spatially heterogeneous and possibly fractal in nature. We model seismicity in spatially heterogeneous stress fields, including the temporal response to static stress perturbations. Starting with 3D models of the crust with fractal-like heterogeneous stress, we add a stress perturbation due to a major earthquake, and couple this system to rate-state seismicity equations to explore the temporal evolution of seismicity during an aftershock sequence. We find that the stress perturbation can generate an increased focal mechanism orientation scatter and a sizable rotation of the average P-T orientation. Both of these effects have been observed in the real Earth. Previous models (without 3D stress heterogeneity) assumed these effects reflected "real" changes in the stress orientation, and hence required small background differential stresses of 10 MPa or less to satisfy the observations. In our model, the presence of stress heterogeneity can bias the failures, creating an apparent stress rotation much larger than the actual stress rotation. Consequently, our model can generate rotations similar to that observed in the real Earth with background differential stresses in the range of 20-50 MPa. We also explore how the focal mechanism orientations rotate and scatter throughout the aftershock cycle. Woessner (2005), observed a step increase in focal mechanism orientation scatter at the onset of a major earthquake with a decay back to smaller scatter during the aftershock sequence. Using rate and state friction seismicity rate equations, we model this temporal evolution of seismicity in our 3D heterogeneous stress volumes.