Will there be an at least 8.0 magnitude earthquake in California before 2028?

San Andreas fault large rupture recurrence intervals vary by location. Paleoseismic studies show that recurrence baselines for significant events, typically M7.5 or greater, are not uniform across the entire fault, rather they are representative records for large events ^[1][2][3]. Over the past 1500 years, mean recurrence intervals have been approximately 105 years at Wrightwood and 145 years on average at Pallett Creek ^[1][2][3]. More recent analyses suggest that the recurrence intervals for large (M7.5+) events might range from 145 to 200 years ^[1][2][3].

Current southern San Andreas quiescence exceeds historical mean intervals. As of May 6, 2026, the southern San Andreas segment has experienced a quiescent period longer than the paleoseismic mean intervals of 100 to 150 years ^[4][1][2]. The last major rupture, the 1857 Fort Tejon event, occurred approximately 169 years ago ^[4]. This extended period indicates the fault is considered “overdue” when measured against average recurrence times at locations like Wrightwood and Pallett Creek, though it has not yet surpassed the complete observed range of historical intervals ^[4][1][2].

Direct comparisons for M8.0+ rupture potential before 2028 are unavailable. Directly computed probabilities for M8.0+ ruptures specifically before 2028 are not available for either the southern San Andreas Fault or the Cascadia Subduction Zone within the retrieved research. Therefore, a direct comparison for this precise timeframe cannot be made based on the provided information, requiring any such assessment to infer relative likelihoods from published 30- or 50-year probabilities ^[2][1][3][4].

Southern San Andreas Fault has a 7% chance of M>=8.0 in 30 years. For the southern San Andreas Fault, materials from the Third Uniform California Earthquake Rupture Forecast (UCERF3) indicate that the likelihood of a California M>=8.0 earthquake within the next 30 years increased to approximately 7%. The UCERF3 model also identifies this fault segment as the most probable source for a large earthquake ^[1][2].

Cascadia Subduction Zone shows higher long-term M>=8.0 rupture probabilities. Regarding the Cascadia Subduction Zone, a 2025 paper published in PNAS reports a time-independent probability of an M>=8 rupture as 15% within the next 50 years ^[3]. Other hazard interpretations from communications by the U.S. Geological Survey (USGS) and Pacific Northwest sources cite a broader range, indicating up to approximately 37% for an M>=8.0 event within 50 years for Cascadia ^[4][5].

Seismologists view foreshocks as potential precursory signs, not standalone predictors. While considered useful for monitoring precursory activity, seismologists from the USGS and Caltech do not consider foreshocks guaranteed indicators of an impending M8.0+ event ^[3][4][5]. The USGS defines foreshocks as earthquakes that precede larger mainshocks in the same location, but they are specifically identified only after the larger mainshock has occurred ^[3]. Caltech research, analyzing data from 2008–2017 in southern California, revealed that 72% of M4+ mainshocks were preceded by foreshock activity significantly elevated above local background levels. These sequences typically lasted from several days to weeks, with a median duration of 16.6 days ^[1]. Despite this, the USGS explicitly frames foreshocks as identifiable only after the mainshock, and discussions of intermediate-term phenomena emphasize probabilistic increased likelihood rather than certainty for prediction ^[3][4][5].

Aseismic slip offers monitoring utility, but not guaranteed M8+ prediction. The USGS notes that active faulting frequently involves a combination of seismic and aseismic behavior, including episodic slow slip and tremor, as well as geodetically or observationally detectable aseismic slip occurring concurrently with foreshock activity ^[5]. The USGS suggests that aseismic deformation can be valuable for monitoring precursory activity ^[5]. For instance, a Caltech study inferred that aseismic slip persisted for more than two months following mainshocks, imposing stresses that helped trigger subsequent earthquakes ^[6]. Additionally, the USGS documented "intermediate-term" pre-earthquake signals in California from 1975–1986, which included the sudden appearance of earthquakes in previously inactive areas and described surface fault creep as a sensitive indicator of stress changes ^[4]. Caltech-linked research identified 92 long-duration earthquake swarms in southern California from 2008–2020, where 53% showed ultra-slow diffusive patterns consistent with fluid injection processes, implying ongoing transient aseismic drivers ^[2]. Nevertheless, these sources consistently describe aseismic deformation as potentially useful for monitoring, but not a standalone guaranteed indicator of an impending M8+ event ^[5].

California's faults require real-time monitoring for seismic assessment. Real-time data is crucial for assessing stress changes and crustal deformation on California's major fault systems, supporting seismic hazard assessment, earthquake early warning systems, and ongoing research into earthquake physics and crustal deformation ^[4][1][2]. This comprehensive monitoring involves several types of instruments, including Global Positioning System (GPS) data, creepmeters, strainmeters, and tiltmeters.

GPS and creepmeters provide critical fault slip and deformation insights. Daily and near real-time GPS data are collected from extensive networks across the western U.S., including high-rate networks in the San Francisco Bay Area and Long Valley. This data provides critical insights into ground velocities, time series, and maps, allowing scientists to detect the relative motion between tectonic plates and understand fault slip ^[1][2][5]. Creepmeters are specifically designed to monitor "fault creep," which is the slow, continuous, and aseismic surface slip on certain faults. These instruments offer micron-precision measurements and are primarily located in areas such as the San Francisco Bay Area (Hayward, Calaveras, San Andreas faults) and near San Juan Bautista and Parkfield, with changes in creep rates indicating alterations in the shear strain applied to a fault ^{[6][3][4][7][8]}.

Sensitive instruments like strainmeters and tiltmeters monitor subtle crustal changes. Highly sensitive strainmeters measure minute changes in crustal strain and are installed deep underground to minimize surface noise, providing continuous data on crustal deformation ^[4][9][10]. These networks are strategically placed along active faults, including the San Andreas and Hayward faults, and in volcanically active regions like Long Valley, to monitor changes associated with fault slip, earthquakes, and volcanic activity ^[10]. Tiltmeters, often deployed alongside strainmeters and creepmeters, measure changes in the tilt of the ground, contributing to a comprehensive understanding of crustal deformation ^[4][10]. Additionally, InSAR data aids in observing various deformation patterns, with its results validated by GPS monitoring data, while the USGS operates extensive seismic networks for rapid earthquake location and magnitude determination ^[11][12].

UCERF3 forecasts M8.0+ events using time-dependent elastic-rebound logic. This model applies elastic-rebound/renewal time-dependent logic to forecast M8.0+ seismic events specifically before 2028, incorporating magnitude-dependent aperiodicity and managing faults with unknown historic open intervals ^[1][2][3][4]. The framework accounts for epistemic uncertainty through a logic-tree set, which it represents by sampling 1,440 alternative logic-tree branches ^[1][2][3][4]. This process considers alternative deformation models, a new smoothed seismicity algorithm, different total rate values for M>=5 events, and various magnitude-scaling relationships as the most influential uncertainties affecting its forecasts ^[3][4].

UCERF3 has significant limitations, including inability to model triggering processes. Key limitations of the UCERF3 framework include its inability to model earthquake-triggering processes or aftershocks ^[5][6]. The model itself cautions that it serves as an approximation and that its sampled range of models is limited, for instance, being constrained to remain similar to UCERF2 ^[3][4]. Consequently, M8.0+ probabilities, particularly for the nearer future before 2028, inherit these structural constraints and are sensitive to the choices made within the model ^[3][4][1]. The predicted probability for a rare event like an M8.0+ can also shift substantially because UCERF3-TD employs renewal-model probability updates and logic-tree epistemic uncertainty, involving 5,760 forecast combinations depending on which branches dominate for the particular evaluation metric and region ^[1][2].

The Uniform California Earthquake Rupture Forecast, Version 3 (UCERF3) estimates California s probability of an M8.0+ earthquake in the next 30 years at about 7.0% (statewide), implying that the chance over only ~2.7 years (from 2026-05-06 to 2028-12-31) would be much smaller than 7% and therefore the event is more likely than not to NOT occur before 2028-12-31 ^[1].

UCERF3 is the long-term forecast produced by USGS, California Geological Survey, Southern California Earthquake Center, and partners, and it specifically provides estimates for chances of large earthquakes over several decades ^[1]. A Kalshi market exists for  at least 8.0 magnitude earthquake in California with a deadline before Jan 1, 2027 (and other similar horizons exist), indicating that prediction-market participants are trying to quantify this low-frequency, high-impact tail risk ^[2][3].