Assessing seismic hazards in active regions necessitates mapping capable faults and determining their recurrence intervals. While surface-rupturing faults are readily identifiable through geologic and geomorphic features, blind faults pose a significant challenge due to their subsurface nature. The 2010 Pichilemu earthquake (Mw 7.0) in Chile, rupturing an unknown fault, underscores this challenge. This event, along with others like the 2016 Kaikoura earthquake (New Zealand) and the 2019 Ridgecrest sequence (California), exemplifies the limitations in our understanding of seismogenesis from hidden faults. Traditional fault mapping relies on surface ruptures and deformation; however, in blind fault settings, this evidence is absent or ambiguous. Blind faults are common in sedimentary basins, often identified through geophysical imaging or indirect geomorphic observations, but estimating their seismic potential remains difficult. This study focuses on the Pichilemu Fault (PIF), a blind fault revealed by the 2010 earthquake, using off-fault geomorphic strain markers to quantify deformation and estimate its seismic potential. The study's context lies in the changing stress distribution in subduction zones, where megathrust earthquakes trigger upper-plate extension, leading to increased seismicity and potential reactivation of crustal faults. While some crustal faults exhibit frequent reactivation, others show recurrence times in the millennia, posing significant local hazards due to the potential for higher amplitude seismic waves and tsunami amplification compared to megathrust events. This study aims to quantify the long-term slip rate of the PIF, infer its earthquake recurrence rate, and investigate the relationship between megathrust earthquakes and PIF slip to better understand the seismic hazard presented by cryptic faults along subduction zones.

Previous research highlights the significant challenge posed by blind faults in seismic hazard assessment. Studies on the 2010 Darfield and 2011 Christchurch earthquakes in New Zealand, the 2016 Kaikoura earthquake, and the 2019 Ridgecrest events in California demonstrated that large-magnitude earthquakes can rupture previously unmapped or incompletely mapped faults. The common approach of mapping active faults based on geomorphic and geologic surface features proves ineffective for blind faults. While geophysical imaging and indirect geomorphic observations offer some insights, estimating the seismic potential of these hidden structures remains problematic. The impact of megathrust earthquakes on upper-plate stress fields and the resulting triggering of crustal earthquakes has been observed in various subduction zones, including Japan, Alaska, and Chile. The recurrence times of these triggered events vary considerably, ranging from decades to millennia depending on the slip rate and characteristics of the upper-plate faults. The location and characteristics of upper-plate faults along subduction zone coastlines represent a significant seismic hazard due to their potential to generate large local seismic waves and tsunami amplification effects. Existing literature establishes the need for comprehensive mapping of these structures and quantifying their slip rates in relation to megathrust earthquake cycles.

The study combines multiple methodologies to investigate the Pichilemu Fault (PIF). First, the seismotectonic and geologic setting was established, noting the PIF's location in the central Chilean margin, a region of significant tectonic activity where the Nazca Plate subducts beneath the South American Plate. The area comprises the Coastal Range, primarily composed of crystalline metamorphic rocks, sculpted by uplifted marine terraces. The dense vegetation and soil cover presented challenges in mapping geological structures. The study then analyzes on-fault tectonic geomorphology, using a high-resolution LiDAR digital terrain model (DTM) to analyze fluvial metrics and the surface expression of the PIF. Two catchments (C1 and C2) were studied, analyzing catchment asymmetry, local relief, slope, drainage networks, and knickpoints. The lack of clear fault scarps suggested the absence of recent surface-rupturing events, implying the fault's blind nature. Next, off-fault tectonic geomorphology was investigated using uplifted marine terraces as strain markers. The terraces were mapped using LiDAR DTM and TerraceM-2 software, with post-infrared infrared stimulated luminescence (post-IR IRSL) dating used to determine their ages. The age of six distinct sedimentary units, corresponding to wave-built terraces, was determined. The spatial distribution and elevations of these terraces reveal a warping pattern consistent with uplift and tilting associated with the PIF. Co-seismic slip and long-term slip rates were then estimated. The co-seismic slip of the 2010 earthquakes was estimated by combining data from GPS and InSAR with aftershock seismicity to define fault geometries. Forward elastic dislocation modeling was used to search for the parameters that best-fit the InSAR and GPS data. To check consistency, results from the forward model were compared with an inverse model. The long-term slip rate was estimated by forward modeling the spatial pattern of uplift rates from marine terraces, creating two models (one using shoreline angles and the other using interpolated uplift rates). Finally, earthquake recurrence was determined by combining the probability distributions of long-term and co-seismic slip rates. Coulomb failure stress (ACFS) modeling was employed to explore potential megathrust earthquake triggering scenarios, evaluating how different megathrust slip distributions impact stress on the PIF. The origin of the PIF and implications for blind faults were discussed, considering the fault's occurrence within a rheologically heterogeneous upper crust.

The study's key findings revolve around the quantification of the PIF's slip rate and recurrence interval and the influence of megathrust earthquakes on its activity. The analysis of off-fault deformed marine terraces provided a long-term slip rate of 0.52 ± 0.04 m/ka. Integrating this with coseismic deformation patterns from the 2010 earthquakes and assuming characteristic slip behavior, the study inferred a recurrence time of 2.12 ± 0.2 ka for Mw 7.0 normal-faulting earthquakes. The analysis of on-fault geomorphic features revealed subtle deformation patterns but no clear evidence of surface rupture, supporting the fault's blind nature. Forward modeling of coseismic displacements using InSAR and GPS data, constrained by aftershock seismicity, indicated a maximum coseismic slip of 1.1 m for fault branch F1 and 0.1 m for branch F2 during the 2010 events. The consistency of these results was checked against an inverse model that also indicated predominantly blind faulting. The long-term deformation model of marine terraces, well-matched with a model based on interpolated uplift rates, suggested similar slip rates as the coseismic model, confirming the long-term pattern of deformation. The combination of long-term slip rate and coseismic slip resulted in a recurrence interval estimation. Coulomb failure stress modeling showed that different slip distributions during megathrust earthquakes could either trigger or inhibit slip on the PIF, with shallow northern megathrust slip promoting PIF slip. The study suggests that the PIF's long recurrence time is due to sporadic triggering during megathrust events, with only a fraction of these events leading to PIF reactivation. Furthermore, the study proposed that the PIF's blind nature might be a consequence of rheological heterogeneities in the upper crust, preventing localized surface faulting. The comparison of the PIF with the El Yolki fault highlighted the role of crustal rheology in controlling the surface expression of faults.

The findings address the research question by quantifying the long-term slip rate and recurrence time of the previously unknown Pichilemu blind fault. The estimated 2.12 ka recurrence time for Mw 7 earthquakes is significantly longer than the recurrence interval of megathrust earthquakes in the region (0.1-0.2 ka). This suggests that the PIF is sporadically triggered by megathrust events, primarily when slip occurs within the shallower northern portion of the megathrust. The Coulomb stress modeling supports this hypothesis, showing that slip on the PIF is favored by stress changes associated with shallow rupture segments. The study highlights the importance of considering the complex interplay between megathrust earthquakes and upper-plate structures in seismic hazard assessments. The rheological heterogeneity of the upper crust appears to play a crucial role in controlling the fault's blind nature. The results are relevant to the broader field of seismic hazard assessment, demonstrating the significance of cryptic faults in subduction zones and the importance of utilizing off-fault geomorphic data for their identification and characterization.

This study significantly advances our understanding of blind faults and their seismic potential. By integrating geomorphic, geodetic, and seismological data with numerical modeling, the study quantified the long-term slip rate and recurrence time of the Pichilemu Fault, a previously unknown structure. The findings emphasize the importance of considering the complex interactions between megathrust earthquakes and upper-plate faults in seismic hazard assessments. Future research could focus on investigating other similar cryptic faults, refining the models to incorporate more sophisticated rheological properties, and expanding the analysis to cover a wider range of subduction zone settings. A more detailed investigation into the crustal rheology around the PIF could shed further light on the factors controlling its blind nature.

The study's limitations primarily relate to the complexity of modeling fault behavior and the inherent uncertainties associated with the employed methods. The models used simplified assumptions regarding fault geometry and rheological properties. While two models (based on shoreline angles and interpolated rates) were used for consistency checks, variations in the results could indicate uncertainties arising from the complex nature of fault slip. The spatial resolution of the data might not capture all the nuances of surface deformation. The long-term slip rate estimation depends on the accuracy of the marine terrace age estimations and correlations with global sea level curves. The Coulomb stress modeling relies on simplifying assumptions about megathrust rupture patterns and frictional properties.