skip to main content


The NSF Public Access Repository (NSF-PAR) system and access will be unavailable from 10:00 PM ET on Friday, December 8 until 2:00 AM ET on Saturday, December 9 due to maintenance. We apologize for the inconvenience.

This content will become publicly available on January 18, 2024

Title: Question-Driven Ensembles of Flexible ETAS Models
Abstract The development of new earthquake forecasting models is often motivated by one of the following complementary goals: to gain new insights into the governing physics and to produce improved forecasts quantified by objective metrics. Often, one comes at the cost of the other. Here, we propose a question-driven ensemble (QDE) modeling approach to address both goals. We first describe flexible epidemic-type aftershock sequence (ETAS) models in which we relax the assumptions of parametrically defined aftershock productivity and background earthquake rates during model calibration. Instead, both productivity and background rates are calibrated with data such that their variability is optimally represented by the model. Then we consider 64 QDE models in pseudoprospective forecasting experiments for southern California and Italy. QDE models are constructed by combining model parameters of different ingredient models, in which the rules for how to combine parameters are defined by questions about the future seismicity. The QDE models can be interpreted as models that address different questions with different ingredient models. We find that certain models best address the same issues in both regions, and that QDE models can substantially outperform the standard ETAS and all ingredient models. The best performing QDE model is obtained through the combination of models allowing flexible background seismicity and flexible aftershock productivity, respectively, in which the former parameterizes the spatial distribution of background earthquakes and the partitioning of seismicity into background events and aftershocks, and the latter is used to parameterize the spatiotemporal occurrence of aftershocks.  more » « less
Award ID(s):
Author(s) / Creator(s):
; ; ; ;
Date Published:
Journal Name:
Seismological Research Letters
Page Range / eLocation ID:
829 to 843
Medium: X
Sponsoring Org:
National Science Foundation
More Like this

    Earthquakes come in clusters formed of mostly aftershock sequences, swarms and occasional foreshock sequences. This clustering is thought to result either from stress transfer among faults, a process referred to as cascading, or from transient loading by aseismic slip (pre-slip, afterslip or slow slip events). The ETAS statistical model is often used to quantify the fraction of clustering due to stress transfer and to assess the eventual need for aseismic slip to explain foreshocks or swarms. Another popular model of clustering relies on the earthquake nucleation model derived from experimental rate-and-state friction. According to this model, earthquakes cluster because they are time-advanced by the stress change imparted by the mainshock. This model ignores stress interactions among aftershocks and cannot explain foreshocks or swarms in the absence of transient loading. Here, we analyse foreshock, swarm and aftershock sequences resulting from cascades in a Discrete Fault Network model governed by rate-and-state friction. We show that the model produces realistic swarms, foreshocks and aftershocks. The Omori law, characterizing the temporal decay of aftershocks, emerges in all simulations independently of the assumed initial condition. In our simulations, the Omori law results from the earthquake nucleation process due to rate and state friction and from the heterogeneous stress changes due to the coseismic stress transfers. By contrast, the inverse Omori law, which characterizes the accelerating rate of foreshocks, emerges only in the simulations with a dense enough fault system. A high-density complex fault zone favours fault interactions and the emergence of an accelerating sequence of foreshocks. Seismicity catalogues generated with our discrete fault network model can generally be fitted with the ETAS model but with some material differences. In the discrete fault network simulations, fault interactions are weaker in aftershock sequences because they occur in a broader zone of lower fault density and because of the depletion of critically stressed faults. The productivity of the cascading process is, therefore, significantly higher in foreshocks than in aftershocks if fault zone complexity is high. This effect is not captured by the ETAS model of fault interactions. It follows that a foreshock acceleration stronger than expected from ETAS statistics does not necessarily require aseismic slip preceding the mainshock (pre-slip). It can be a manifestation of a cascading process enhanced by the topological properties of the fault network. Similarly, earthquake swarms might not always imply transient loading by aseismic slip, as they can emerge from stress interactions.

    more » « less

    The spatio-temporal properties of seismicity give us incisive insight into the stress state evolution and fault structures of the crust. Empirical models based on self-exciting point processes continue to provide an important tool for analysing seismicity, given the epistemic uncertainty associated with physical models. In particular, the epidemic-type aftershock sequence (ETAS) model acts as a reference model for studying seismicity catalogues. The traditional ETAS model uses simple parametric definitions for the background rate of triggering-independent seismicity. This reduces the effectiveness of the basic ETAS model in modelling the temporally complex seismicity patterns seen in seismic swarms that are dominated by aseismic tectonic processes such as fluid injection rather than aftershock triggering. In order to robustly capture time-varying seismicity rates, we introduce a deep Gaussian process (GP) formulation for the background rate as an extension to ETAS. GPs are a robust non-parametric model for function spaces with covariance structure. By conditioning the length-scale structure of a GP with another GP, we have a deep-GP: a probabilistic, hierarchical model that automatically tunes its structure to match data constraints. We show how the deep-GP-ETAS model can be efficiently sampled by making use of a Metropolis-within-Gibbs scheme, taking advantage of the branching process formulation of ETAS and a stochastic partial differential equation (SPDE) approximation for Matérn GPs. We illustrate our method using synthetic examples, and show that the deep-GP-ETAS model successfully captures multiscale temporal behaviour in the background forcing rate of seismicity. We then apply the results to two real-data catalogues: the Ridgecrest, CA 2019 July 5 Mw 7.1 event catalogue, showing that deep-GP-ETAS can successfully characterize a classical aftershock sequence; and the 2016–2019 Cahuilla, CA earthquake swarm, which shows two distinct phases of aseismic forcing concordant with a fluid injection-driven initial sequence, arrest of the fluid along a physical barrier and release following the largest Mw 4.4 event of the sequence.

    more » « less
  3. Abstract

    Foreshocks are the only currently widely identified precursory seismic behavior, yet their utility and even identifiability are problematic, in part because of extreme variation in behavior. Here, we establish some global trends that help identify the expected frequency of foreshocks as well the type of earthquake most prone to foreshocks. We establish these tendencies using the global earthquake catalog of the U.S. Geological Survey National Earthquake Information Center with a completeness level of magnitude 5 and mainshocks with Mw≥7.0. Foreshocks are identified using three clustering algorithms to address the challenge of distinguishing foreshocks from background activity. The methods give a range of 15%–43% of large mainshocks having at least one foreshock but a narrower range of 13%–26% having at least one foreshock with magnitude within two units of the mainshock magnitude. These observed global foreshock rates are similar to regional values for a completeness level of magnitude 3 using the same detection conditions. The foreshock sequences have distinctive characteristics with the global composite population b-values being lower for foreshocks than for aftershocks, an attribute that is also manifested in synthetic catalogs computed by epidemic-type aftershock sequences, which intrinsically involves only cascading processes. Focal mechanism similarity of foreshocks relative to mainshocks is more pronounced than for aftershocks. Despite these distinguishing characteristics of foreshock sequences, the conditions that promote high foreshock productivity are similar to those that promote high aftershock productivity. For instance, a modestly higher percentage of interplate mainshocks have foreshocks than intraplate mainshocks, and reverse faulting events slightly more commonly have foreshocks than normal or strike-slip-faulting mainshocks. The western circum-Pacific is prone to having slightly more foreshock activity than the eastern circum-Pacific.

    more » « less

    We propose a theoretical modelling framework for earthquake occurrence and clustering based on a family of invariant Galton–Watson (IGW) stochastic branching processes. The IGW process is a rigorously defined approximation to imprecisely observed or incorrectly estimated earthquake clusters modelled by Galton–Watson branching processes, including the Epidemic Type Aftershock Sequence (ETAS) model. The theory of IGW processes yields explicit distributions for multiple cluster attributes, including magnitude-dependent and magnitude-independent offspring number, cluster size and cluster combinatorial depth. Analysis of the observed seismicity in southern California demonstrates that the IGW model provides a close fit to the observed earthquake clusters. The estimated IGW parameters and derived statistics are robust with respect to the catalogue lower cut-off magnitude. The proposed model facilitates analyses of multiple quantities of seismicity based on self-similar tree attributes, and may be used to assess the proximity of seismicity to criticality.

    more » « less
  5. Abstract

    The number of aftershocks increases with mainshock size following a well‐defined scaling law. However, excursions from the average behavior are common. This variability is particularly concerning for large earthquakes where the number of aftershocks varies by factors of 100 for mainshocks of comparable magnitude. Do observable factors lead to differences in aftershock behavior? We examine aftershock productivity relative to the global average for all mainshocks () from 1990 to 2019. A global map of earthquake productivity highlights the influence of tectonic regimes. Earthquake depth, lithosphere age, and plate boundary type correspond well with earthquake productivity. We investigate the role of mainshock attributes by compiling source dimensions, radiated seismic energy, stress drop, and a measure of slip heterogeneity based on finite‐fault source inversions for the largest earthquakes from 1990 to 2017. On an individual basis, stress drop, normalized rupture width, and aspect ratio most strongly correlate with aftershock productivity. A multivariate analysis shows that a particular set of parameters (dip, lithospheric age, and normalized rupture area) combines well to improve predictions of aftershock productivity on a cross‐validated data set. Our overall analysis is consistent with a model in which the volumetric abundance of nearby stressed faults controls the aftershock productivity rather than variations in source stress. Thus, we suggest a complementary approach to aftershock forecasts based on geological and rupture properties rather than local calibration alone.

    more » « less