S21B-0274 0800h
A Comparative Aftershock Study for the 2003 Armeria, Colima, Mexico, Earthquake (Mw=7.4)
The seismic sequence of the January 22, 2003 Armeria Earthquake, was recorded by the two permanent seismic networks in the region: the Colima Seismic Network (RESCO) and the Jalisco Seismic Network (RESJAL). Additionally a portable seismic network with seven digital stations with triaxial sensors was deployed from 25 to 31 January. A total of 110 aftershocks were located by RESJAL and RESCO for whole the period, and 272 events were located using data from the portable network. Analysis of the first 72 hours (72 events) shows two groups of events, one the near of the 1995 Earthquake (M = 8.0) epicenter with an almost vertical distribution of events between 2 and 12 km depth. The second group around and north of the mainshock with a distribution dipping about $40\deg$ in an almost EW direction which agrees with focal mechanism for mainshock. These data indicate that the Earthquake was a continental intraplate inverse fault event and do not support the possible rupture of the Colima Gap. The aftershocks recorded by the portable network present a similar epicentral distribution, however most of the events are deeper, between 12 and 30 km. We observe two tendencies one seems distributed between 12 and 22 km along the interphase between Jalisco Block and Rivera Plate assuming a $12\deg$ dipping. The second one between 22 and 30 km depth and dipping about $40\deg$ across (?) the Rivera Plate. We present a comparative analysis of these results including join focal mechanism solutions for both sets of data.
S21B-0275 0800h
Tidal Triggering Effect on Earthquake Occurrence Precursory to Large Thrust Earthquakes in Subduction Zones
We observed a clear tidal triggering effect on earthquake occurrence precursory to large thrust earthquakes associated with the subduction of oceanic plates. We measured the correlation between the Earth tide and earthquake occurrence using shallow reverse-fault type earthquakes with Mw $\geq$ 5.0 in and around the focal regions of eleven interplate earthquakes with Mw $\geq$ 7.5. For each earthquake, we assigned the tidal phase angle at the origin time by theoretically calculating the tidal stress change on the fault plane. Based on the distribution of the tidal phase angles, we statistically tested whether they concentrate near some particular angle or not by using the Schuster's test. As a result of analysis, a precursory tidal effect was found for six earthquakes. For all the six cases, a high correlation appeared for several years preceding the occurrence of large earthquakes, and the high correlation disappeared after the main shocks. The frequency distribution of tidal phase angles in the pre-event period exhibited a peak at the phase angle where the tidal stress change is at its maximum to accelerate the fault slip. This indicates that the high correlation observed in the pre-seismic stage is not a stochastic chance but is a physical consequence of the tidal stress change. The optimum value of the frictional coefficient varies in a wide range from 0.2 to infinity (only the normal stress change is effective) showing a positive correlation with the age of subducting oceanic plate.
S21B-0276 0800h
Seasonal Seismicity Associated with Western United States Volcanoes
We examined volcanic areas in the western United States for evidence of "seasonal" seismicity (annual periodicity). Data sets cover a 20-year period and include six distinct zones of focused seismicity in large calderas (3 zones apiece at Long Valley caldera and Yellowstone National Park) and four Cascade Range stratovolcanoes (Mts. Lassen, Hood, St. Helens, and Rainier). We developed a method to identify and remove large swarm events associated with magmatic intrusion from the time series in order to better detect annual trends. We then applied statistical techniques to test for annual seismic patterns. Seven of the ten study areas show statistically significant seasonality ($>$80% probability in Schuster tests). The exceptions are two areas in Long Valley caldera that were dominated by seismic swarms during the majority of the 20-year period and Mt. St. Helens, where the apparent seasonality fails to meet our 80% probability threshold. Elsewhere in the Cascade Range and in the other Long Valley study area, peaks in annual seismicity occur in the late summer and autumn. In Yellowstone National Park, seismicity peaks earlier, during late spring and summer. In order to determine the most plausible cause for the observed trends, we quantify possible forcing mechanisms that could induce seasonal seismicity. Possible triggers include barometric pressure changes, solid earth tides, unloading associated with annual snowmelt, and groundwater recharge. Annual cycles of barometric pressure and solid earth tides cause changes in annual loads of $<$5 kPa, which are probably not sufficient to trigger seismicity. The stress changes associated with snow unloading and groundwater recharge are larger ($>$10 kPa) and may serve as the trigger for seasonal seismicity.
S21B-0277 0800h
Earthquake Triggering and Fault Interaction in a Thrust Fault System: the 1999 Chi-Chi, Taiwan, Earthquake
To understand earthquake triggering and possible fault interaction in a thrust fault system, we calculate the changes in seismicity rate associated with the 1999 Chi-Chi, Taiwan, earthquake, and the Coulomb stress change caused by the earthquake. The Coulomb stress change for the ruptured Chelungpu fault shows large stress changes near the fault, and broad lobes of stress increase and decrease off the fault. The pattern of the aftershock distribution show similar feature to the Coulomb stress change calculated for a simple slip model, suggesting its strong association. More than 70% of the aftershocks in strike-slip and thrust faulting mechanisms can be explained by the increased Coulomb stress changes. Most of the calculated zones of Coulomb stress increase, often by less than 1 bar, correspond to sites of aftershocks. Although the Chi-Chi earthquake is a thrust event, many of the triggered aftershocks show strike-slip mechanisms, particularly to the 2 south and southeast of the southern tip of the thrust fault. The calculated Coulomb stress change on the likely rupture planes of the M>6 aftershocks is also generally positive, with values of 2-5 bars. This suggests that the large aftershocks do not necessarily occur where the Coulomb stress change is most positive. The feature for stress shadow is also observed to the further west of the fault, but less significant to the observation for stress triggering. Thrust faults to the east of the ruptured Chelungpu fault lie in calculated Coulomb stress shadows, whereas the Changhua and Chuko faults, to the west and south of the Chelungpu fault, are subject to stress increases. Comparison with historical events since 1600, the calculated Coulomb stress change shows strong correlation of the occurrences of the earthquake on the Chelungpu fault to the events on Changhua and Chuko faults in the 19th century. This seems to suggest the possible predictive power of the static stress calculation in estimating occurrences of future large earthquakes.
S21B-0278 0800h
Hydraulically Induced Seismicity at the KTB - Event Dislocation Caused by Anisotropy
The environment of the German Continental Deep drilling site (KTB) is known to be anisotropic. In this study we have analyzed the seismic events generated by the KTB long term injection experiment. For a period of 60 days a total of more than 4000 $m^3$ of water were injected into the KTB generating more than 2500 seismic events. About 260 events were recorded not only by the borehole geophone but also by the 40 station surface array centered at the KTB. For this experiment we are in the favorable position to have a priory knowledge, where the acoustic emission should originate, i.e., they should start at the injection point slowly propagating away from it. Localizing the events recorded in the borehole and by the surface array using an isotropic model based on average velocities derived from check shots leads to a southward lateral shift of the center of the event cloud of about 500m away from the injection point. Since the total extension of the elongated event cluster is about 2 km in E-W direction and about 300-400 m in N-S direction this is a significant contribution. Using the 3-D isotropic velocity model obtained from the KTB reflection data, no significant change in the localization is observed and the lateral shift is reproduced. The application of station corrections is inappropriate since the geophones of the recording array are located on the same formation. Localizing the same events using an anisotropic model based on published data centers the event cloud around the injection point. The anisotropic localization not only removes the lateral shift of the events but it also significantly alters the shape of the event cloud from an elongated cluster to an almost circular distribution. This is important if the spatio-temporal evolution is interpreted with respect to the hydraulic properties of the rocks. Also interpretations with respect to event clustering due to tectonics or hydro fracing may be severely affected if the localization is based on the wrong model assumptions. The localization of the data with the simple anisotropic model lead to a physically reasonable distribution of events which is not obtained with isotropic models. In seismology most often there is no a priory information on the subsurface available and we are in a less favorable position than for the KTB injection experiment. A dislocation of the events due to anisotropy may be overlooked misleading the interpretation of the distribution of the events and its spatio-temporal evolution.
S21B-0279 0800h
Evaluating Short-Term Probability Forecasts of M4+ Earthquakes in California and Western Nevada Since 2001
Since 2001 we have made a number of forecasts of 10-day time periods of increased probability of earthquakes of M4 or greater (M4+) in California and Nevada. The forecasts are based on observations of non-Poissonian short-term clustering in the M4+ seismicity from 1968 to 2000 for Northern California/Western Nevada and from 1983 to 2000 for Southern California, and they assume that the M4+ earthquake statistics for these earlier time periods can be applied without change to the current and future seismicity. Since 2001, the observed versus expected outcome of the forecasts: for Northern California/Western Nevada has been: 24 forecasts, 10 forecasts expected correct, and 7 correct forecasts. For Central California/Western Nevada it has been: 33 forecasts, 11 forecasts expected correct, 8 correct forecasts. For Southern California it has been: 19 forecasts, 7 forecasts expected correct, 0 correct forecasts. For Northern and Central California and Western Nevada, the difference between the observed and expected number of correct forecasts is statistically insignificant. However, for Southern California the probability of 0 correct forecasts out of 19 trials is only 0.2%. Since 2001, the M4+ seismicity in Southern California shows a strongly regular component with a period between 25-30 days, suggesting a connection to the monthly earth tidal cycle. Our analysis indicates that the temporal pattern of Southern California M4+ seismicity underwent a major change sometime in 2000-2001, even though the average rate of M4+ earthquakes remained constant.
S21B-0280 0800h
Relation Between Mainshock Rupture Process and Omori Law for Aftershock Moment Release Rate
We compare the source time functions (moment release rates) of three large California mainshocks with the seismic moment release rates during their aftershock sequences. Aftershock moment release rates, computed by summing aftershock moments in time intervals, follow the Omori law from minutes to months after the mainshock. Furthermore, in contrast to the previously-observed saturation in numbers of aftershocks shortly after the mainshock rupture, no such saturation is seen in the aftershock moment release rates, which are dominated by the largest aftershocks. We argue that the observed saturation in aftershock numbers, described by the "time offset" parameter c in Omori's law, is likely an artifact due to the under-reporting of small aftershocks, which is related to the difficulty of detecting large numbers of small aftershocks in the mainshock coda. We further propose that it is more natural for c to be negative (i.e., singularity follows the onset of mainshock rupture) than positive (singularity precedes onset of rupture). To make a more general comparison of mainshock rupture process and aftershock moment rates, we then scale mainshock time functions to equalize the effects of the varied seismic moments. For the three California mainshocks, we compare the scaled time functions with similarly-scaled aftershock moment rates. Finally, we compare global averages of scaled time functions of many shallow events to the average scaled aftershock moment release rate for six California mainshocks. In each of the above comparisons, the extrapolation of the aftershock moment rates according to Omori's law back in time toward the mainshock rupture indicates that the temporal intensity of the aftershock moment release is about 1.5 orders of magnitude less than the maximum reached by the mainshock rupture. This may be due to the differing amplitudes and relative importance of static and dynamic stresses in aftershock initiation compared to mainshock rupture propagation.
http://scec.ess.ucla.edu/~ykagan/heidi_index.html
S21B-0281 0800h
Fracturing From Seismic Waves At Regional Distances As Inferred From Water Well Tidal Response.
Water level responses to solid earth tides provide a means for measuring aquifer properties and, by extension, fracturing. At the time of five earthquakes in Southern California, we observe transient changes up to 20$^o$ in the phase of the water level response to the dilational volumetric strain of the $M_2$ tidal component of wells at the Pi\~non Flat Observatory (PFO). Larger, closer earthquakes generate larger phase changes than the smaller ones. The duration of the phase change is between two and ten months after which the response returns to the background value. We use a model of cylindrical flow driven by an imposed head oscillation through a single, laterally extensive, confined, homogeneous and isotropic aquifer to interpret changes in phase response as changes in aquifer properties. We interpret the changes as due to changes in the transmissivity (rate of water transmission through a unit width of aquifer under a unit hydraulic gradient). The transmissivity measures the ease with which water flows through the aquifer and is proportional to permeability. The higher the transmissivity, the smaller the observed tidal phase lag. At the time of the earthquakes, the transmissivity at the Pi\~non Flat Observatory increases by a factor that can be as high as four. Two different simple permeability models are used to interpret changes in transmissivity as changes in the geometry of the aquifer. Both a network of tubes model and a set of plane parallel fractures model give similar results. The transmissivity changes imply increases between 5% and 70% in the diameter of the tubes and the thickness of the fractures. For instance, the 1999 Hector Mine earthquake produces a 30% increase in fracture size for the tube model and a 40% increase for the fracture model. The fracture model always gives higher values for the fracture size change than the tube model. The fraction of average crack size increment shows a direct dependence on the amplitude of seismic waves at PFO as calculated from empirical magnitude relationships.