Geology topics

Andrew J. Michael

Publications and source records attributed to Andrew J. Michael.

At least 37 records · Page 2Linked to original sources

USGS2022

An efficient, analytic solution using order statistics for probabilistic seismic‐hazard assessment without the Poisson assumption

Standard approaches to probabilistic seismic‐hazard assessment (PSHA) assume that earthquakes are random, independent events that follow a Poisson distribution of occurrences in a given time period ( Cornell, 1968 ). To overcome the limitations of the Poisson assumption, such as ignoring earthquake clustering, we introduce an analytic method for PSHA that uses order statistics to allow for arbitrary distributions of earthquake occurrence. Cornell (1968) used the Poisson assumption to achieve a computationally efficient method that enables users to explore the impact of parameters used in earthquake occurrence and ground‐motion models. We apply our order statistics method to the highly clustered seismicity associated with caldera collapses at Kīlauea and explore the general implications of non‐Poisson behavior for PSHA. We find that non‐Poisson behavior has the greatest impact for high probabilities of exceedance, low‐mean rates of occurrence, and multiple exceedances. Those conditions can be important for applications such as operating standards for buildings and infrastructure engineering, standards for temporary structures and during construction, the insurance industry, the design of earthquake early warning, and to assess hazards due to clustered processes such as aftershock sequences and earthquake swarms. The commonly used rate of exceedance hides the difference between the hazard due to a non‐Poisson distribution and a Poisson distribution with the same mean rate of earthquakes. Thus, including non‐Poisson behavior in PSHA means that we must plot and discuss PSHA results as the probability and not the rate of exceedance.

Bulletin of the Seismological Society of America↗

USGS2022

Prospective and retrospective evaluation of the U.S. Geological Survey public aftershock forecast for the 2019-2021 Southwest Puerto Rico Earthquake and aftershocks

The M w "> M w 6.4 Southwest Puerto Rico Earthquake of 7 January 2020 was accompanied by a robust fore‐ and aftershock sequence. The U.S. Geological Survey (USGS) has issued regular aftershock forecasts for more than a year since the mainshock, available on a public webpage. Forecasts were accompanied by interpretive and informational material, published in English and Spanish. Informational products included narrative “scenarios” for how the aftershock sequence could play out, infographics, and a report on the potential duration of the aftershock sequence through the next decade. Forecasts are based on the epidemic‐type aftershock sequence (ETAS) model and generated using the USGS AftershockForecaster software—an interactive graphical user interface built on the OpenSHA platform ( Field et al. , 2003 ). The initial forecast is based on past sequences in similar tectonic environments; subsequent forecasts are tuned to the ongoing sequence via Bayesian model updating. Probabilistic aftershock forecasts for the next day, week, month, and year were publicly released and archived at a daily to monthly tempo, allowing for a truly prospective test of the forecast. Here, we evaluate the forecast over the first year of the recorded aftershocks. The ETAS‐based forecast performed well overall, successfully capturing both the chance of having at least one earthquake of a given magnitude in a forecast interval as well as the non‐Poissonian distribution of the total number of aftershocks within an interval. A retrospective analysis shows that the ETAS model is a substantial improvement over the existing Reasenberg and Jones (1989) forecast model. The exercise also reveals some limitations of the current model, in particular, with respect to nonstationarities in the aftershock magnitude distribution and model parameters throughout the evolving sequence.

Puerto Rico↗

USGS2022

Modeling the occurrence of M ∼ 5 caldera collapse-related earthquakes in Kīlauea volcano, Hawai'i

During the 2018 Kīlauea eruption and caldera collapse, M ∼ 5 caldera collapse earthquakes occurred almost daily from mid-May until the beginning of August. While caldera collapses happen infrequently, the collapse-related seismicity damaged nearby structures, and so these events should be included in a complete seismic hazard assessment. Here, we present an approach to forecast the seismic hazard of the collapse earthquakes. We model their occurrence by combining a Poisson distribution for the number of collapses with a negative binomial for the number of earthquakes in a collapse, based on observations at Kīlauea. This rate model is then combined with a ground motion model to assess the seismic hazard posed by caldera collapse events. The rate model is non-Poisson but a Poisson model is adequate for low exceedance probabilities (e.g., <10% in 50 years). This approach could be generalized to model the hazard from earthquakes triggered by other underlying processes.

Hawaii↗

USGS2021

2021 U.S. National Seismic Hazard Model for the State of Hawaii

The 2021 U.S. National Seismic Hazard Model (NSHM) for the State of Hawaii updates the two-decades-old former model by incorporating new data and modeling techniques to improve the underlying ground shaking forecasts of tectonic-fault, tectonic-flexure, volcanic, and caldera collapse earthquakes. Two earthquake ground shaking hazard models (public policy and research) are produced that differ in how they account for declustered catalogs. The earthquake source model is based on (1) declustered earthquake catalogs smoothed with adaptive smoothing methods, (2) earthquake rate forecasts based on three temporally varying 60-year time periods, (3) maximum magnitude models that extend to larger earthquakes than previously considered, (4) a separate Kīlauea-specific seismogenic caldera collapse model which accounts for clustered event behavior observed during the 2018 eruption, and (5) fault ruptures that consider historic seismicity, GPS-based strain rates, and a new Quaternary fault database. Two new Hawaii-specific ground motion models (GMMs) and five additional global models consistent with Hawaii shaking data are used to forecast ground shaking at 23 spectral periods and peak parameters. Site effects are modeled using western U.S. and Hawaii specific empirical equations and provide shaking forecasts for eight site classes. For most sites the new model results in similar spectral accelerations as those in the 2001 NSHM, with a few exceptions caused mostly by GMM changes. Ground motions are highest in the southern portion of the Island of Hawai‘i due to high rates of forecasted earthquakes on décollement faults. Shaking decays to the northwest where lower earthquake rates result from flexure of the tectonic plate. Large epistemic uncertainties in source characterizations and GMMs lead to an overall high uncertainty (more than a factor of 3) in ground shaking at Honolulu and Hilo. The new shaking model indicates significant chances of slight or greater damaging ground motions across most of the island chain.

Hawai'i↗

USGS2021

An International Virtual Workshop on Global Seismology and Tectonics (IVWGST‐2020)

An International Virtual Workshop on Global Seismology and Tectonics (IVWGST‐2020) was organized by the Geoscience and Technology Division of Council of Scientific and Industrial Research—North East Institute of Science and Technology, Jorhat, India from 14 to 25 September 2020. This workshop predominantly catered to undergraduate, postgraduate, and Ph.D. students, scientists, and academicians from across the globe. The primary motive of IVWGST-2020 was to inspire the participating students, perturbed by the unprecedented situation brought about by the COVID-19 pandemic, with quality lecture sessions, so as to lift their spirits. The virtual workshop served as a conduit for the students and researchers to directly interact with several pioneers and prominent geoscience researchers from around the world. Lectures, via Microsoft Teams, were given by 15 eminent speakers from diverse geoscience forums and institutions, and were attended by more than 1000 participants, mostly students and researchers, from 30 different countries. This report briefly summarizes the agenda, describes our experiences hosting the virtual workshop, and documents the challenges faced.

Seismological Research Letters↗

USGS2020

The normal faulting 2020 Mw5.8 Lone Pine, Eastern California earthquake sequence

The 2020 M w "> M w 5.8 Lone Pine earthquake, the largest earthquake on the Owens Valley fault zone, eastern California, since the nineteenth century, ruptured an extensional stepover in that fault. Owens Valley separates two normal‐faulting regimes, the western margin of the Great basin and the eastern margin of the Sierra Nevada, forming a complex seismotectonic zone, and a possible nascent plate boundary. Foreshocks began on 22 June 2020; the largest M w "> M w 4.7 foreshock occurred at ∼ 6    km "> ∼6 km depth, with primarily normal faulting, followed ∼ 40    hr "> ∼ 40 hr later on 24 June 2020 by an M w "> M w 5.8 mainshock at ∼ 7    km "> ∼ 7 km depth. The sequence caused overlapping ruptures across a ∼ 0.25    km 2 "> ∼0.25 km 2 area, extended to ∼ 4    km 2 "> ∼ 4 km 2 , and culminated in an ∼ 25    km 2 "> ∼25 km 2 aftershock area. The mainshock was predominantly normal faulting, with a strike of 330° (north‐northwest), dipping 60°–65° to the east‐northeast. Comparison of background seismicity and 2020 Ridgecrest aftershock rates showed that this earthquake was not an aftershock of the Ridgecrest mainshock. The M w "> M w M w – m B "> – m B relationship and distribution of ground motions suggest typical rupture speeds. The aftershocks form a north‐northwest‐trending, north‐northeast‐dipping, 5 km long distribution, consistent with the rupture length estimated from analysis of regional waveform data. No surface rupture was reported along the 1872 scarps from the 2020 M w "> M w 5.8 mainshock, although, the dipping rupture zone of the M w "> M w 5.8 mainshock projects to the surface in the general area. The mainshock seismic energy triggered rockfalls at high elevations ( ⁠ > 3.0    km "> > 3.0 km ⁠ ) in the Sierra Nevada, at distances of 8–20 km, and liquefaction along the western edge of Owens Lake. Because there were ∼ 30 % "> ∼ 30 % fewer aftershocks than for an average southern California sequence, the aftershock forecast probabilities were lower than expected. ShakeAlert, the earthquake early warning system, provided first warning within 9.9 s, as well as subsequent updates.

California↗

USGS2020

Pseudo-prospective evaluation of UCERF3-ETAS forecasts during the 2019 Ridgecrest sequence

The 2019 Ridgecrest sequence provides the first opportunity to evaluate Uniform California Earthquake Rupture Forecast v.3 with epidemic‐type aftershock sequences (UCERF3‐ETAS) in a pseudoprospective sense. For comparison, we include a version of the model without explicit faults more closely mimicking traditional ETAS models (UCERF3‐NoFaults). We evaluate the forecasts with new metrics developed within the Collaboratory for the Study of Earthquake Predictability (CSEP). The metrics consider synthetic catalogs simulated by the models rather than synoptic probability maps, thereby relaxing the Poisson assumption of previous CSEP tests. Our approach compares statistics from the synthetic catalogs directly against observations, providing a flexible approach that can account for dependencies and uncertainties encoded in the models. We find that, to the first order, both UCERF3‐ETAS and UCERF3‐NoFaults approximately capture the spatiotemporal evolution of the Ridgecrest sequence, adding to the growing body of evidence that ETAS models can be informative forecasting tools. However, we also find that both models mildly overpredict the seismicity rate, on average, aggregated over the evaluation period. More severe testing indicates the overpredictions occur too often for observations to be statistically indistinguishable from the model. Magnitude tests indicate that the models do not include enough variability in forecasted magnitude‐number distributions to match the data. Spatial tests highlight discrepancies between the forecasts and observations, but the greatest differences between the two models appear when aftershocks occur on modeled UCERF3‐ETAS faults. Therefore, any predictability associated with embedding earthquake triggering on the (modeled) fault network may only crystalize during the presumably rare sequences with aftershocks on these faults. Accounting for uncertainty in the model parameters could improve test results during future experiments.

California↗

USGS2020

Regionally Optimized Background Earthquake Rates from ETAS (ROBERE) for probabilistic seismic hazard assessment

We use an epidemic‐type aftershock sequence (ETAS) based approach to develop a regionally optimized background earthquake rates from ETAS (ROBERE) method for probabilistic seismic hazard assessment. ROBERE fits parameters to the full seismicity catalog for a region with maximum‐likelihood estimation, including uncertainty. It then averages the earthquake rates over a suite of catalogs from which foreshocks and aftershocks have been removed using stochastic declustering while maintaining the same Gaussian smoothing currently used for the U.S. Geological Survey National Seismic Hazard Model (NSHM). The NSHM currently determines these rates by smoothing a single catalog from which foreshocks and aftershocks have been removed using the method of Gardner and Knopoff (1974 ; hereafter, GK74). The parameters used in GK74 were determined from subjectively identified aftershock sequences, unlike ROBERE, in which both background rate and aftershock triggering parameters are objectively fitted. A major difference between the impacts of the two methods is GK74 significantly reduces the b "> b ‐value, a critical value for seismic hazard analysis, whereas ROBERE maintains the original b "> b ‐value from the full catalog. We apply these methods to the induced seismicity in Oklahoma and Kansas and tectonic activity in the San Francisco Bay Region. Using GK74 gives lower overall earthquake rates but estimates higher hazard due to the reduction in the b "> b ‐value. ROBERE provides higher earthquake rates, at the magnitude of completeness, but lower hazard because it does not alter the b "> b ‐value. We test two other declustering methods that produce results closer to ROBERE but do not use objectively fit parameters, include uncertainty, and may not work as well in other areas. We suggest adopting ROBERE for the NSHM so that our hazard estimates are based on an objective analysis, including uncertainty, and do not depend strongly on potentially biased b "> b ‐values, which was never the goal of the existing methodology.

Bulletin of the Seismological Society of America↗

USGS2020

Potential duration of aftershocks of the 2020 southwestern Puerto Rico earthquake

Abstract Aftershocks (earthquakes clustered spatially and chronologically near the occurrence of a causative earthquake) are ongoing in southwestern Puerto Rico after a series of earthquakes, which include a magnitude 6.4 earthquake that occurred near Barrio Indios, Guayanilla, on January 7, 2020, and affected the surrounding area. This report estimates the expected duration of these aftershocks by incorporating observations of aftershocks as of January 17, 2020, into a well-established statistical model of how earthquake sequences behave. Aftershocks will persist for years to decades, although with decreasing frequency, and earthquakes will likely be felt on a daily basis for up to several months. These estimates have significant uncertainty owing to different scenarios of how the earthquake sequence may evolve over time and could also change if a new large aftershock occurs. This report also estimates the amount of time remaining until the annual probability of magnitude 5, 6, and 7 or greater aftershocks—which could cause additional damage—decreases to 50, 25, 10, 5, and 1 percent. As of this writing, the chance of having a magnitude 6 or greater earthquake within a given year, going forward, will not fall below 25 percent for another 3 months to 3 years. The chance of having a magnitude 5 or greater earthquake will not fall below 25 percent for a decade or more. The aftershocks discussed in this report would be located in the same general area as the aftershocks that have already occurred. Our results do not imply a change in the risk of earthquakes in other parts of Puerto Rico. Resumen Las réplicas (terremotos agrupados espacial y cronológicamente cerca de la ocurrencia de un terremoto causante) están en curso en el suroeste de Puerto Rico después de una serie de terremotos, que incluyen un terremoto de magnitud 6.4, ocurrido cerca del Barrio Indios, Guayanilla, el 7 de enero de 2020 y que afectaron las áreas circundantes. Este informe estima la duración esperada de las réplicas incorporando observaciones de réplicas a partir del 17 de enero de 2020 en un modelo estadístico bien establecido de cómo se comportan las secuencias de terremotos. Las réplicas persistirán durante años o décadas, aunque con una frecuencia decreciente, y los terremotos probablemente se sentirán a diario durante varios meses. Estos estimados tienen una incertidumbre significativa debido a diferentes escenarios de cómo la secuencia del terremoto puede evolucionar con el tiempo y también podrían cambiar si ocurre una nueva réplica grande. Este informe también estima la cantidad de tiempo restante hasta que la probabilidad anual de réplicas de magnitud 5, 6 y 7 o más - que podría causar daños adicionales- disminuya a un 50, 25, 10, 5 y 1 por ciento. Al momento de escribir este artículo, la posibilidad de tener un terremoto de magnitud 6 en un año determinado, en el futuro, no caerá por debajo del 25 por ciento durante otros 3 meses a 3 años. La probabilidad de tener un terremoto de magnitud 5 o mayor no será inferior al 25 por ciento durante una década o más. Las réplicas discutidas en este informe se ubicarían en la misma área general que las réplicas que ya han ocurrido. Nuestros resultados no implican un cambio en el riesgo de terremotos en otras partes de Puerto Rico.

Puerto Rico↗

USGS2020

Global seismology and tectonics: A report on International Virtual Workshop on Global Seismology and Tectonics (IVWGST-2020)

No abstract available.

Current Science↗

USGS2019

Ensembles of ETAS models provide optimal operational earthquake forecasting during swarms: Insights from the 2015 San Ramon, California swarm

Earthquake swarms, typically modeled as time-varying changes in background seismicity that are driven by external processes such as fluid flow or aseismic creep, present challenges for operational earthquake forecasting. While the time decay of aftershock sequences can be estimated with the modified Omori law, it is difficult to forecast the temporal behavior of seismicity rates during a swarm. To explore these issues, we apply the Epidemic-Type Aftershock Sequence (ETAS) model (Ogata, JASA, 1988) to the 2015 San Ramon, California swarm, which lasted several weeks and had almost 100 2≤M≤3.6 earthquakes. We develop 3-day forecasts during the swarm based on an ETAS model fit to all prior seismicity in the region as well as an ETAS model fit only to previous swarms in the region, which is better at capturing the higher background rate during the swarm. We also explore forecasts where the background rate is updated periodically during the swarm using data over different lookback windows and find that generally these models perform better than the models where the background rate is fixed. Finally, we construct ensemble forecasts by combining the different models weighted according to their performance. The ensemble forecasts outperform all of the individual models and allow us to avoid making arbitrary choices at the outset of a swarm as to which single model will perform the best.

California↗

USGS2019

Developing earthquake forecast templates for fast and effective communication

No abstract available.

Conference Paper↗

USGS2018

On the potential duration of the aftershock sequence of the 2018 Anchorage earthquake

Currently, an aftershock sequence is ongoing in Alaska after the magnitude 7.0 Anchorage earthquake of November 30, 2018. Using two scenarios, determined with observations as of December 14, 2018, this report estimates that it will take between 2.5 years and 3 decades before the rate of aftershocks decays to the rate of earthquakes that were occurring in this area before the magnitude 7.0 mainshock. All of the time estimates have significant uncertainty owing to different scenarios of how the sequence may decrease over time and could also change if a large aftershock occurs. The report also estimates the amount of time after the mainshock until the annual probability of magnitude 5 or greater and 6 or greater aftershocks—which could cause additional damage—decreases to 50, 25, 10, and 5 percent. For instance, the probability of one or more magnitude 6 or greater aftershocks in the following year decreases to 10 percent between 7 and 250 days after the mainshock. The same probability for magnitude 5 or greater earthquakes is reached between 500 and 7,000 days after the mainshock.

Alaska↗

USGS2018

Updated California aftershock parameters

Reasenberg and Jones (1989) introduced a statistical model for aftershock rate following a mainshock along with estimates of “generic” California parameter values based on past aftershock sequences. The Reasenberg and Jones (1989) model has been used for decades to issue aftershock forecasts following M ≥ 5 "> M ≥ 5 mainshocks in California. Here, we update the “generic” parameters for California through a fit to the aftershock sequences of M ≥ 5 "> M ≥ 5 mainshocks occurring since 1980. We find aftershock productivity values that are lower on average than the generic productivity reported by Reasenberg and Jones (1989) , likely because low‐productivity sequences were omitted from their analysis and possibly because of a trade‐off between productivity and b "> b ‐value. We confirm the observation of Llenos and Michael (2017) that southern California sequences are more productive on average than northern California sequences. The Mendocino area is much less productive; the hydrothermal areas in Long Valley, Coso, and the Salton Sea, in contrast, are much more productive. We also quantify the variability of the Reasenberg and Jones (1989) productivity parameter a "> a a between sequences with a normal distribution. This distribution of a "> a ‐values can be used to compute aftershock forecasts that include epistemic uncertainty and can be used as the prior for Bayesian updating of the a "> a a ‐value as a sequence progresses.

California↗

USGS2018

Preface to the Focus Section on the Collaboratory for the Study of Earthquake Predictability (CSEP): New results and future directions

The Collaboratory for the Study of Earthquake Predictability (CSEP; Jordan, 2006 ) carries out fully prospective tests of earthquake forecasts, using fixed and standardized statistical tests and authoritative data sets, to assess the predictive skill of forecast models and to make objective comparisons between models. CSEP conducts prospective experiments at four testing centers around the world, at which more than 400 models and model versions are currently under evaluation. These models include a range of methods and scales from long‐term global earthquake forecasts to short‐term regional forecasts used for Operational Earthquake Forecasting (OEF). CSEP has also conducted retrospective tests and developed new testing methods in its quest to answer fundamental scientific questions, improve seismic hazard assessments, and develop new forecast methods for OEF.

Seismological Research Letters↗

USGS2018

2018 one‐year seismic hazard forecast for the central and eastern United States from induced and natural earthquakes

This article describes the U.S. Geological Survey (USGS) 2018 one‐year probabilistic seismic hazard forecast for the central and eastern United States from induced and natural earthquakes. For consistency, the updated 2018 forecast is developed using the same probabilistic seismicity‐based methodology as applied in the two previous forecasts. Rates of earthquakes across the United States M ≥ 3.0 "> M ≥ 3.0 grew rapidly between 2008 and 2015 but have steadily declined over the past 3 years, especially in areas of Oklahoma and southern Kansas where fluid injection has decreased. The seismicity pattern in 2017 was complex with earthquakes more spatially dispersed than in the previous years. Some areas of west‐central Oklahoma experienced increased activity rates where industrial activity increased. Earthquake rates in Oklahoma (429 earthquakes of M ≥ 3 "> M ≥ 3 and 4 M ≥ 4 "> M ≥ 4 ), Raton basin (Colorado/New Mexico border, six earthquakes M ≥ 3 "> M ≥ 3 ), and the New Madrid seismic zone (11 earthquakes M ≥ 3 "> M ≥ 3 ) continue to be higher than historical levels. Almost all of these earthquakes occurred within the highest hazard regions of the 2017 forecast. Even though rates declined over the past 3 years, the short‐term hazard for damaging ground shaking across much of Oklahoma remains at high levels due to continuing high rates of smaller earthquakes that are still hundreds of times higher than at any time in the state’s history. Fine details and variability between the 2016–2018 forecasts are obscured by significant uncertainties in the input model. These short‐term hazard levels are similar to active regions in California. During 2017, M ≥ 3 "> M ≥ 3 earthquakes also occurred in or near Ohio, West Virginia, Missouri, Kentucky, Tennessee, Arkansas, Illinois, Oklahoma, Kansas, Colorado, New Mexico, Utah, and Wyoming.

Seismological Research Letters↗

USGS2017

Forecasting the (un)productivity of the 2014 M 6.0 South Napa aftershock sequence

The 24 August 2014 M w 6.0 South Napa mainshock produced fewer aftershocks than expected for a California earthquake of its magnitude. In the first 4.5 days, only 59 M ≥1.8 aftershocks occurred, the largest of which was an M 3.9 that happened a little over two days after the mainshock. We investigate the aftershock productivity of the South Napa sequence and compare it with other M ≥5.5 California strike‐slip mainshock–aftershock sequences. While the productivity of the South Napa sequence is among the lowest, northern California mainshocks generally have fewer aftershocks than mainshocks further south, although the productivities vary widely in both regions. An epidemic‐type aftershock sequence (ETAS) model ( Ogata, 1988 ) fit to Napa seismicity from 1980 to 23 August 2014 fits the sequence well and suggests that low‐productivity sequences are typical of this area. Utilizing regional variations in productivity could improve operational earthquake forecasting (OEF) by improving the model used immediately after the mainshock. We show this by comparing the daily rate of M ≥2 aftershocks to forecasts made with the generic California model ( Reasenberg and Jones, 1989 ; hereafter, RJ89), RJ89 models with productivity updated daily, a generic California ETAS model, an ETAS model based on premainshock seismicity, and ETAS models updated daily following the mainshock. RJ89 models for which only the productivity is updated provide better forecasts than the generic RJ89 California model, and the Napa‐specific ETAS models forecast the aftershock rates more accurately than either generic model. Therefore, forecasts that use localized initial parameters and that rapidly update the productivity may be better for OEF than using a generic model and/or updating all parameters.

California↗

USGS2017

A synoptic view of the Third Uniform California Earthquake Rupture Forecast (UCERF3)

Probabilistic forecasting of earthquake‐producing fault ruptures informs all major decisions aimed at reducing seismic risk and improving earthquake resilience. Earthquake forecasting models rely on two scales of hazard evolution: long‐term (decades to centuries) probabilities of fault rupture, constrained by stress renewal statistics, and short‐term (hours to years) probabilities of distributed seismicity, constrained by earthquake‐clustering statistics. Comprehensive datasets on both hazard scales have been integrated into the Uniform California Earthquake Rupture Forecast, Version 3 (UCERF3). UCERF3 is the first model to provide self‐consistent rupture probabilities over forecasting intervals from less than an hour to more than a century, and it is the first capable of evaluating the short‐term hazards that result from multievent sequences of complex faulting. This article gives an overview of UCERF3, illustrates the short‐term probabilities with aftershock scenarios, and draws some valuable scientific conclusions from the modeling results. In particular, seismic, geologic, and geodetic data, when combined in the UCERF3 framework, reject two types of fault‐based models: long‐term forecasts constrained to have local Gutenberg–Richter scaling, and short‐term forecasts that lack stress relaxation by elastic rebound.

Seismological Research Letters↗