FORECASTING EARTHQUAKES WITH THE SAGE MODEL

Earthquake Research and Analysis - New Advances in Seismology, 2013

Journal of African Earth Sciences, 2017

Pure and Applied Geophysics, 2017

A common characteristic of deep mines in hard rock is induced seismicity. This results from stress changes and rock failure around mining excavations. Following large seismic events, there is an increase in the levels of seismicity, which gradually decay with time. Restricting access to areas of a mine for enough time to allow this decay of seismic events is the main approach in re-entry strategies. The statistical properties of aftershock sequences can be studied with three scaling relations: (1) Gutenberg-Richter frequency magnitude, (2) the modified Omori's law (MOL) for the temporal decay, and (3) Båth's law for the magnitude of the largest aftershock. In this paper, these three scaling relations, in addition to the stochastic Reasenberg-Jones model are applied to study the characteristic parameters of 11 large magnitude mininginduced aftershock sequences in four mines in Ontario, Canada. To provide guidelines for re-entry protocol development, the dependence of the scaling relation parameters on the magnitude of the main event are studied. Some relations between the parameters and the magnitude of the main event are found. Using these relationships and the scaling relations, a space-time-magnitude re-entry protocol is developed. These findings provide a first approximation to concise and well-justified guidelines for re-entry protocol development applicable to the range of mining conditions found in Ontario, Canada.

Tectonophysics, 2013

This study assesses the aftershock activity of two earthquakes that occurred on December 20, 2010 with magnitude of M N 6.5 (Global CMT M w 6.5) and January 27, 2011 with magnitude of M N 6.0 (Global CMT M w 6.2) in the Rigan region of southeastern Iran. This study has been done by assessing the statistical properties of the aftershock sequences associated with each of these earthquakes, namely b-value of GutenbergeRichter relation, partitioning of radiated seismic energy, p-value of modified Omori law and the D C-value associated with the fractal dimension. The b-values of b ¼ 0.89 ± 0.08 and b ¼ 0.88 ± 0.08 were calculated for first main shock and second main shock sequence respectively. This suggests that this region is characterized by large differential stress; the genesis of large aftershock activity in a short time interval gives power this. Further, 2.2% of the whole energy is related with the aftershocks activity for first main shock sequence while 97.8% is associated with main shock; for second sequence, 20% of the total energy is associated with the aftershocks activity while 80% is associated with main shock. The pvalues of 1.1 ± 0.12 and 1.1 ± 0.1 were calculated for first and second main shocks sequence respectively, which imply fast decay rate of aftershocks and high surface heat flux. A value of the spatial fractal dimension (D c) equal to 2.34 ± 0.03 and 2.54 ± 0.02 for first and second main shocks sequence respectively, which reveals random spatial distribution and source in a two-dimensional plane that is being filled-up by fractures. Moreover, we then use the models to calculate the Coulomb stress change to appraise coming seismic hazard by inspecting the static Coulomb stress field due to these two main shocks for the recognition of the conceivable regions of aftershocks activity. The first main shock increased stress by more than 0.866 bars at the hypocenter of the second main shock, thus promoting the failure. In addition, the cumulative coseismic Coulomb stress changes due to the reveals that most of the aftershocks happened in the region of increased Coulomb stress.

Tectonophysics, 2016

In this research, a new algorithm for generating a stochastic earthquake catalog is presented. The algorithm is based on the acceptance-rejection sampling of von Neumann. The result is a computer simulation of earthquakes based on the calculated statistical properties of each zone. Vere-Jones states that an earthquake sequence can be modeled as a series of random events. This is the model used in the proposed simulation. Contrariwise, Utsu indicates that the mainshocks are special geophysical events. The algorithm has been applied to zones of Chile, China, Spain, Japan, and the USA. This allows classifying the zones according to Vere-Jones' or Utsu's model. The results have been quantified relating the mainshock with the largest aftershock within the next 5 days (which has been named as Bath event). The results show that some zones fit Utsu's model and others Vere-Jones'. Finally, the fraction of seismic events that satisfy certain properties of magnitude and occurrence is analyzed.

Bulletin of the Seismological Society of America, 2017

We describe a simple method to determine the probability distribution function of the magnitude M max and return period T R of the maximum plausible earthquake on crustal faults. The method requires the background seismicity rate (estimated from instrumental data) and the rate of interseismic moment buildup. The method assumes that the moment released by the seismic slip is in balance with the moment deficit accumulated in between earthquakes. It also assumes that the seismicity obeys the Gutenberg-Richter (GR) law up to M max. We took into account the aftershocks of large infrequent events that were not represented in the instrumental record, so that we could estimate the average seismicity rate over the entire fault history. We extrapolated the instrumental record, using the GR law to model the frequency of larger events and their aftershocks. This increased the frequencies of smaller events on average; when these smaller events were newly extrapolated, they predicted a higher frequency of larger events. We iterated this process until stability was reached, and then we assumed moment balance when we found the maximum magnitude; we have found this method to be appropriate in applications involving examples of fault with good historical catalogs. We then showed examples of applications to faults with no historical catalogs. We present results from nine cases. For the San Andreas fault system, we find M

Journal of Geophysical Research, 2010

Any earthquake can trigger more earthquakes. This triggering occurs in both the classical aftershock zone as well as the far field. These populations of triggered earthquakes may or may not be distinct in terms of triggering mechanism. Here we look for a distinction between the populations by examining how the observed intensity of triggering scales with the amplitude of the triggering strain in each. To do so, we apply a new statistical metric based on earthquake interevent times to a large data set and measure earthquake triggering as a function of dynamic strain amplitude, where strain is estimated from empirical ground motion regressions. This method allows us to identify triggering at dynamic strain amplitudes down to 3 × 10 -9 , orders of magnitude smaller than previously reported. This threshold appears to be an observational limit and shows that extremely small dynamic strains can trigger faults that are sufficiently near failure. Using a probabilistic model to transform measured interevent times to seismicity rate changes, we find that triggering rates in the far field scale with peak dynamic strain. This scaling, projected into the near field, accounts for 15%-60% of earthquakes within 6 km of magnitude 3-5.5 earthquakes. Statistical seismicity simulations validate the interevent time method and show that the data are consistent with the number of far-field triggered earthquakes being linearly proportional to peak dynamic strain. We interpret the additional near-field component as reflecting either static stress triggering, more effective dynamic triggering at higher frequencies, or a concentration of aftershock nucleation sites very near main shocks.

Bulletin of the Seismological Society of America, 2013

Large megathrust subduction earthquakes generate prolific aftershock sequences which last over an extended period of time and affect wide spatial areas. Among those aftershocks, the largest ones can cause additional damage and pose significant risks to population and infrastructure. Therefore, modeling aftershock sequences of large subduction earthquakes is of considerable importance for seismic-hazard assessment and earthquake risk mitigation. It can also play a prominent role in the ground-shaking modeling of major mainshock-aftershock sequences. In this work, we analyze statistical properties of aftershock sequences of large subduction earthquakes worldwide which occurred from 1973 to present, including recent catastrophic events in Sumatra, Chile, and Japan. We use information provided in the National Earthquake Information Center (NEIC) catalog to extract 70 aftershock sequences generated by mainshocks of M 7.0 and above. We construct their temporal decay rates and magnitudefrequency statistics. To model their temporal behavior, we estimate the parameters of the modified Omori law. In the magnitude domain, we model the frequency-magnitude statistics using the Gutenberg-Richter scaling relationship. We also analyze statistically the difference between the magnitude of the mainshock and the corresponding largest aftershock in the sequence and discuss this in terms of Båth's law. One of the main goals of this work is to investigate the variation in parameter values of the above empirical laws with respect to the magnitude of the mainshock. Our main finding indicates that most parameters do not depend on the magnitude of the mainshock. However, they show some variation in values across different subduction settings.

Journal of Asian Earth Sciences, 2012

A damaging and widely felt moderate earthquake (Mw 6.4) hit the rural, mountainous region of southwestern Pakistan on October 28, 2008. The main shock was followed by another earthquake of identical magnitude (Mw 6.4) on the next day. The spatial distribution of aftershocks and focal mechanism revealed a NW-SE striking rupture with right-lateral strike-slip motion which is sympathetic to the NNW-SSE striking active mapped Urghargai Fault. The occurrence of strike-slip earthquakes suggests that along with the thrust faults, strike slip faults too are present beneath the fold-and-thrust belt of Sulaiman-Kirthar ranges and accommodates some of the relative motion of the Indian and Eurasian plates.

Pure and Applied Geophysics, 2017

This study considers the statistics of fluid-induced and remotely triggered seismicity at The Geysers geothermal field, California. Little seismicity was reported before steam extraction began in 1960. Beginning in 1980 the residual water associated with power generation was re-injected, producing induced seismicity. Beginning in 1997 large-scale injections of cold water began to enhance the generation of steam. This led to an increase in M \ 1.2 earthquakes from approximately 5 per month to 20. Two excellent seismic networks generate two earthquake catalogs for the fluid-induced seismicity at The Geysers. Although this seismicity satisfies Gutenberg-Richter (GR) scaling to a good approximation, the scaling parameters differ. We propose a correction that eliminates this problem. We show that the seismicity at The Geysers is nearly independent of time for the period 2009-2014 and suggest that this supports our hypothesis that the seismic moment release is in near balance with the geodetic moment accumulation in the region. Our study demonstrates that aftershocks of the larger fluid-induced earthquakes also satisfy GR scaling as well as Omori's law for their time dependence. Our results support the hypothesis that the earthquakes are caused by the reduction in friction on faults due to the injected fluids. Statistics of remotely triggered earthquakes and their associated aftershocks at The Geysers are also presented. The 8/24/ 14 M = 6.02 South Napa earthquake triggered an M & 4.38 event as well as some 80 other M [ 1.25 events. The GR and decay statistics are given. However, to separate aftershocks from remotely triggered earthquakes, an additional triggered sequence is studied. The M = 7.2 4/4/10 Baja earthquake triggered some 34 M [ 1.25 earthquakes at The Geysers in the first hour including an M = 3.37 event. We conclude that the remotely triggered seismicity is dominated by local aftershocks of the larger remotely triggered earthquakes.