ALBERT

Description: [1]  Seismicity closely related to hydrological impacts has been observed in several locations worldwide; particularly in intraplate areas where tectonic stressing rates are small. The triggering mechanism is usually explained by a poroelastic response of the seismogenic crust to surface water flux, leading to pore pressure changes at depth. To explain the earthquake triggering in response of those small stress changes, however, the crust has to be near a critical state in which other transient processes might be significant. One of the prominent examples is the Mt. Hochstaufen in SW Germany, where seismicity is known to vary seasonally. A previous analysis showed that the seismicity in 2002 was highly correlated with model forecasts based on fluid diffusion and rate- and state-dependent frictional nucleation. Here we revisit this case by accounting additionally for poroelastic effects, as well as for thermoelastic and tidal stresses. We also test whether the model can explain the observations of the subsequent eight years between 2003 and 2010. Our analysis confirms that rainfall is the dominant driving force in this region. The model not only fits the year 2002 activity very well, but provides with the same parameters a reasonable fit to the subsequent period, with a probability gain of about 4 per event in comparison to a time-independent Poisson model.

Description: The West Bohemia/Vogtland region, central Europe, is well-known for its repeating swarm activity. However, the latest activity in 2014, although spatially overlapping with previous swarm activity, consisted of three classical aftershock sequences triggered by M L 3.5, 4.4, and 3.4 events. To decode the apparent system change from swarm-type to mainshock-aftershock characteristics, we have analyzed the details of the major M L 4.4 sequence based on focal mechanisms and relocated earthquake data. Our analysis shows that the mainshock occurred with rotated mechanism in a step-over region of the fault plane, unfavorably oriented to the regional stress field. Most of its intense aftershock activity occurred in-plane with classical characteristics such as (i) the maximum magnitude of the aftershocks is significantly less than the mainshock magnitude and (ii) the decay can be well fitted by the Omori-Utsu law. However, the absolute number of aftershocks and the fitted Omori-Utsu c and p parameters are much larger than for typical sequences. By means of the epidemic type aftershock sequence (ETAS) model, we show that an additional aseismic source with an exponentially decaying strength triggered a large fraction of the aftershocks. Corresponding pore pressure simulations with an exponentially decreasing flow rate of the fluid source show a good agreement with the observed spatial migration front of the aftershocks extending approximately with log( t ). Thus we conclude that the mainshock opened fluid pathways from a finite fluid source into the fault plane explaining the unusual high rate of aftershocks, the migration patterns, and the exponential decrease of the aseismic signal.

Description: Abstract The aftershock productivity is known to strongly vary for different mainshocks of the same magnitude, which cannot be simply explained by random fluctuations. In addition to variable source mechanisms, different rheological properties might be responsible for the observed variations. Here we show, for the subduction zone of northern Chile, that the aftershock productivity is linearly related to the degree of mechanical coupling along the subduction interface. Using the earthquake catalog of Sippl et al. (2018, https://doi.org/10.1002/2017JB015384), which consists of more than 100,000 events between 2007 and 2014, and three different coupling maps inferred from interseismic geodetic deformation data, we show that the observed aftershock numbers are significantly lower than expected from the Båth's law. Furthermore, the productivity decays systematically with depth in the uppermost 80 km, while the b value increases. We show that this lack of aftershocks and the observed depth dependence can be simply explained by a linear relationship between the productivity and the coupling coefficient, leading to Båth law only in the case of full coupling. Our results indicate that coupling maps might be useful to forecast aftershock productivity and vice versa.

Description: [1]  [7] finds three differences of the seismicity clustering in southern California compared to self-similar triggering models: (i) a significantly lower b -value for the aftershocks; (ii) a too large aftershock number, and (iii) a too large foreshock-aftershock ratio to be consistent with the Båth law. Based on these observations, the author concluded that the observed seismicity is not in agreement with self-similarity triggering and/or the observed clustering is not primarily caused by earthquake-to-earthquake triggering. However, I show that the observed lower b -value is likely related to incomplete recordings after mainshocks and that the apparently too large aftershock number does not disprove the self-similarity. Thus only the enhanced foreshock-to-aftershock ratio seems to indicate some discrepancy to self-similar triggering.

Description: Several mechanisms are proposed to underlie earthquake triggering including static stress interactions and dynamic stress transfer. Significant differences of these mechanisms are particularly expected in the spatial distribution of aftershocks. However, testing the different hypotheses is challenging because it requires the consideration of the large uncertainties involved in stress calculations as well as the appropriate consideration of secondary aftershock triggering which is related to stress changes induced by smaller pre- and aftershocks. In order to evaluate the forecast capability of different mechanisms, we take the effect of smaller-magnitude earthquakes into account by using the epidemic type aftershock sequence (ETAS) model where the spatial probability distribution of direct aftershocks, if available, is correlated to alternative source information and mechanisms. We test surface shaking, rupture geometry, and slip distributions. As an approximation of the shaking level, we use ShakeMap data which are available in near real-time after a main shock and thus could be used for first-order forecasts of the spatial aftershock distribution. Alternatively, we test the use of empirical decay laws related to minimum fault distance and Coulomb stress change calculations based on published and random slip models. For comparison, we analyze the likelihood values of the different model combinations in the case of several well-known aftershock sequences (1992 Landers, 1999 Hector Mine, 2004 Parkfield). Our test shows that the fault geometry is the most valuable information for improving aftershock forecasts. Furthermore, we find that static stress maps can additionally improve the forecasts of off-fault aftershock locations, while the integration of ground shaking data could not upgrade the results significantly.

Description: We study changes in effective stress (normal stress minus pore pressure) that occurred in the French Alps during the 2003–2004 Ubaye earthquake swarm. Two complementary data sets are used. First, a set of 974 relocated events allows us to finely characterize the shape of the seismogenic area and the spatial migration of seismicity during the crisis. Relocations are performed by a double-difference algorithm. We compute differences in travel times at stations both from absolute picking times and from cross-correlation delays of multiplets. The resulting catalog reveals a swarm alignment along a single planar structure striking N130°E and dipping 80°W. This relocated activity displays migration properties consistent with a triggering by a diffusive fluid overpressure front. This observation argues in favor of a deep-seated fluid circulation responsible for a significant part of the seismic activity in Ubaye. Second, we analyze time series of earthquake detections at a single seismological station located just above the swarm. This time series forms a dense chronicle of +16,000 events. We use it to estimate the history of effective stress changes during this sequence. For this purpose we model the rate of events by a stochastic epidemic-type aftershock sequence model with a nonstationary background seismic rate λ0(t). This background rate is estimated in discrete time windows. Window lengths are determined optimally according to a new change-point method on the basis of the interevent times distribution. We propose that background events are triggered directly by a transient fluid circulation at depth. Then, using rate-and-state constitutive friction laws, we estimate changes in effective stress for the observed rate of background events. We assume that changes in effective stress occurred under constant shear stressing rate conditions. We finally obtain a maximum change in effective stress close to −8 MPa, which corresponds to a maximum fluid overpressure of about 8 MPa under constant normal stress conditions. This estimate is in good agreement with values obtained from numerical modeling of fluid flow at depth, or with direct measurements reported from fluid injection experiments.

Description: We perform a retrospective forecast experiment on the 1992 Landers sequence comparing the predictive power of commonly used model frameworks for short-term earthquake forecasting. We compare a modified short-term earthquake probability (STEP) model, six realizations of the epidemic-type aftershock sequence (ETAS) model, and four models that combine Coulomb stress changes calculations and rate-and-state theory to generate seismicity rates (CRS models). We perform the experiment under the premise of a controlled environment with predefined conditions for the testing region and data for all modelers. We evaluate the forecasts with likelihood tests to analyze spatial consistency and the total amount of forecasted events versus observed data. We find that (1) 9 of the 11 models perform superior compared to a simple reference model, (2) ETAS models forecast the spatial evolution of seismicity best and perform best in the entire test suite, (3) the modified STEP model matches best the total number of events, (4) CRS models can only compete with empirical statistical models by introducing stochasticity in these models considering uncertainties in the finite-fault source model, and (5) resolving Coulomb stress changes on 3-D optimally oriented planes is more adequate for forecasting purposes than using the specified receiver fault concept. We conclude that statistical models perform generally better than the tested physics-based models and parameter value updates using the occurrence of aftershocks generally improve the predictive power in particular for the purely statistical models in space and time.

Description: [1]  The empirical Båth's law indicates that the earthquake process is self-similar and provides an opportunity to estimate the magnitude of the largest aftershock subsequent to a mainshock. However, the analysis of this relation is limited to a small magnitude range and also depends on the aftershock selection rules. As an alternative, we analyze in this paper, the cumulative seismic moment of aftershocks relative to the mainshock moment, because (i) it is a physical quantity that does not only take the largest aftershock into account; (ii) background activity can be considered and as a result estimations are less affected by selection rules; and (iii) the effects of the catalog cut-off magnitude can be corrected, what leads to larger magnitude range for the analysis. We analyze the global PDE USGS catalog (combined with CMT focal mechanisms) and find that the seismic moment release of aftershocks is on average approximately 5% of the mainshock seismic moment. We show that the results can be well fitted by simulations of the ETAS model. In particular, we test whether simulations constrained by predictions of the static stress-triggering model, proposing a break of self-similarity due to the finite seismogenic width, are in agreement with observations. Our analysis shows that the observed dependency on the mainshock magnitude as well as systematic variations with the mainshock fault plane solution can be both explained by the constraints based on the static stress-triggering.