Smooth Fault Research Articles

Fault-network geometry influences earthquake frictional behaviour.

Understanding the factors governing the stability of fault slip is a crucial problem in fault mechanics1-3. The importance of fault geometry and roughness on fault-slip behaviour has been highlighted in recent lab experiments4-7 and numerical models8-11, and emerging evidence suggests that large-scale complexities in fault networks have a vital role in the fault-rupture process12-18. Here we present a new perspective on fault creep by investigating the link between fault-network geometry and surface creep rates in California, USA. Our analysis reveals that fault groups exhibiting creeping behaviour show smaller misalignment in their fault-network geometry. The observation indicates that the surface fault traces of creeping regions tend to be simple, whereas locked regions tend to be more complex. We propose that the presence of complex fault-network geometries results in geometric locking that promotes stick-slip behaviour characterized by earthquakes, whereas simpler geometries facilitate smooth fault creep. Our findings challenge traditional hypotheses on the physical origins of fault creep explained primarily in terms of fault friction19-21 and demonstrate the potential for a new framework in which large-scale earthquake frictional behaviour is determined by a combination of geometric factors and rheological yielding properties.

Injection-induced fault slip and associated seismicity in the lab: Insights from source mechanisms, local stress states and fault geometry

Probing source mechanisms of natural and induced earthquakes is a powerful tool to unveil associated rupture kinematics. The source processes of failure and slip instability driven by stress loading are affected by fault geometry, but the source ruptures of injection-induced seismicity in relation to fault structures and local stress states remain poorly understood. We have conducted a series of fault reactivation and slip experiments on sandstone samples containing faults with different surface roughness (smooth saw-cut fault and fractured rough fault). We impose progressive fluid injection to induce fault slip, and simultaneously monitor the associated acoustic emission (AE) activity. Using high-resolution AE recordings, we perform full moment tensor inversion of all located AE sources, and investigate the changes of AE source characteristics associated with induced fault slip and their relation to fault roughness. For the complex and rough fault, we observe significant non-double-couple components of AE sources and a high degree of focal mechanism heterogeneity. The temporal changes of AE mechanisms associated with injection-induced fault slip on the smooth fault reveal increasing proportions of double-couple components and decreasing variability of AE focal mechanisms when approaching the onset of slip events. The observed inconsistency between the nodal planes of AE sources and the macroscopic fault plane orientation is attributed to the development of secondary fracture networks surrounding the principal slip surface. We analyze changes in the magnitude-frequency characteristics and source mechanisms of AEs with fault-normal distance, showing that for the smooth (mature) fault, Gutenberg–Richter b-value of on-fault seismicity is lower and focal mechanisms are less heterogeneous, compared to off-fault seismicity. Our results emphasize the important role of roughness-related changes in local fault geometry and associated stress heterogeneity for source mechanisms and rupture kinematics of injection-induced seismicity.

Major California faults are smooth across multiple scales at seismogenic depth

Surface traces of earthquake faults are complex and segmented on multiple scales. At seismogenic depth the detailed geometry of faults and earthquake rupture is mainly constrained by earthquake locations. Standard earthquake locations are usually too diffuse to constrain multi-scale fault geometry, while differential-timing relocation mainly improves finest scale precision. NLL-SSST-coherence, an enhanced, absolute-timing earthquake location procedure, iteratively generates traveltime corrections to improve multi-scale precision and uses waveform similarity to improve fine-scale precision. Here we apply NLL-SSST-coherence to large-earthquake sequences and background seismicity along strike-slip faults in California. Our relocated seismicity at seismogenic depth along major fault segments and around large-earthquake ruptures often defines smooth, planar or arcuate, near-vertical surfaces across the sub-km to 10’s of km scales. These results show that multi-scale smooth fault segments are characteristic of major, strike-slip fault zones and may be essential to large earthquake rupture. Our results suggest that smoothness and curvature of faults influences earthquake initiation, rupture, rupture direction and arrest, and can define earthquake hazard. The results corroborate that surface traces of strike-slip fault zones reflect complex, shallow deformation and not directly simpler, main slip surfaces at depth, and support use of planar or smoothly curved faults for modeling primary earthquake rupture.

Fault surface morphology as an indicator for earthquake nucleation potential

AbstractLaboratory measurements can determine the potential for geologic materials to generate unstable (seismic) slip, but a direct relation between sliding behavior in the laboratory and physical characteristics observable in the field is lacking, especially for the phyllosilicate-rich gouges that are widely observed in natural faults. We integrated laboratory friction experiments with surface topography microscopy and demonstrated a quantitative correlation between frictional slip behavior and fault surface morphology of centimeter-scale samples. Our results show that striated, smooth fault surfaces were formed in experiments that exhibited stable sliding, whereas potentially unstable sliding was associated with rougher, isotropic fault surfaces. We interpret that frictional stability and fault surface morphology are linked via the evolution of asperity contacts on localized slip surfaces. If fault surface roughness obeys a fractal relationship over a large range of length scales, then we infer that the morphological characteristics observed in the laboratory could indicate the earthquake nucleation potential on natural fault surfaces.

Experimental investigation of the effects of loading rate, contact roughness, and normal stress on the stick-slip behavior of faults

The tectonic movement rate, in situ stresses (e.g., normal stress) and fault roughness may change during the long evolutionary histories of seismogenic faults. To investigate the effects of these three factors on fault slip behaviors, we conducted direct shear tests by sequentially changing the shear loading rate and normal stress on small-scale granite specimens under three roughness conditions. We found that for a given fault system, the shear loading rate is a primary controlling parameter of the recurrence intervals of stick-slip events, and the correlation between them can be empirically expressed by a power law. Stick-slip recurrence intervals can influence the stress drops of stick-slip events via two different tendencies. The shear stress drops increase quasilinearly with the recurrence intervals at a given shear loading rate, while the combined data from all shear rate tests show that the shear stress drops increase with the logarithm of the recurrence time. In the investigated range of fault roughness, we observed that surface roughness conditions strikingly affect the sliding mode and recurrence frequency of stick-slip events. In detail, the two smooth fault systems slide with regular stick-slip events, and the roughest fault system exhibits bimodal behavior. Namely, at lower normal stresses, stick-slip events cannot be stably generated, a few were occasionally generated, and regular stick-slip events were interrupted by local stick-slip. When the applied normal stress is sufficiently high, the fault slides with continuous regular stick-slip events in the same way as those in the smoother fault systems. We further noticed that the recurrence frequency of stick-slip and friction drop are correlated at a fixed sliding distance, and this relationship can be described by an inverse proportional function. Our results suggest that all three factors, namely, shear loading rate, fault surface roughness and normal stress, can influence the behaviors of fault slip and hence are important for evaluating earthquake hazards.

A Study on the Energy Reduction Measures of Data Centers through Chilled Water Temperature Control and Water-Side Economizer

The degree of integration of IT devices and consumption of cooling energy are consistently increasing owing to developments in the data center industry. Hence, to ensure the smooth operation and fault prevention of IT devices, the energy consumption of cooling systems has increased, leading to active research on improvements in cooling system performance for reducing energy consumption. This study examines the reduction in cooling energy consumption using a simulation by applying chilled water control and a water-side economizer (WSE) system to enhance the cooling system efficiency. The simulation results showed that the energy consumption was reduced by 1.8% when the chilled water temperature was set to 11 °C in a conventional system and by up to 19.6% when WSE was also applied. Furthermore, when the changes in chilled water temperature were applied for efficient operation of WSE, the energy consumption was reduced by up to 30.1% compared to that in conventional energy systems.

Stress drop–magnitude dependence of acoustic emissions during laboratory stick-slip

SUMMARYEarthquake source parameters such as seismic stress drop and corner frequency are observed to vary widely, leading to persistent discussion on potential scaling of stress drop and event size. Physical mechanisms that govern stress drop variations are difficult to evaluate in nature and are more readily studied in controlled laboratory experiments. We perform two stick-slip experiments on fractured (rough) and cut (smooth) Westerly granite samples to explore fault roughness effects on acoustic emission (AE) source parameters. We separate large stick-slip events that generally saturate the seismic recording system from populations of smaller AE events which are sensitive to fault stresses prior to slip. AE event populations show many similarities to natural seismicity and may be interpreted as laboratory equivalent of natural microseismic events. We then compare the temporal evolution of mechanical data such as measured stress release during slip to temporal changes in stress drops derived from AEs using the spectral ratio technique. We report on two primary observations: (1) In contrast to most case studies for natural earthquakes, we observe a strong increase in seismic stress drop with AE size. (2) The scaling of stress drop with magnitude is governed by fault roughness, whereby the rough fault shows a more rapid increase of the stress drop–magnitude relation with progressing large stick-slip events than the smooth fault. The overall range of AE sizes on the rough surface is influenced by both the average grain size and the width of the fault core. The magnitudes of the smallest AE events on smooth faults may also be governed by grain size. However, AEs significantly grow beyond peak roughness and the width of the fault core. Our laboratory tests highlight that source parameters vary substantially in the presence of fault zone heterogeneity (i.e. roughness and narrow grain size distribution), which may affect seismic energy partitioning and static stress drops of small and large AE events.

Fault Stability Across the Seismogenic Zone

AbstractAcross the seismogenic zone, the transition from brittle to plastic deformation corresponds to a semibrittle regime where brittle fracturing and plastic flow coexist at high strength conditions. Thorough experimental investigations on brittle‐plastic transition are crucial to understand why natural faults behave in stable or unstable ways at varying crustal depths and why large earthquakes generally nucleate at the base of the seismogenic zone. To investigate semibrittle deformation in carbonates and the conditions promoting it, we reported here the results of experiments performed on Carrara marble saw cut faults in triaxial conditions. We studied the influence of the confining pressure (ranging between 45 and 180 MPa), axial loading rates (0.01 and 1 μm s−1) and initial fault roughness (smooth and rough) on fault (in‐)stability across the brittle‐plastic transition. We conclude that laboratory earthquakes may nucleate on inherited fault interfaces at brittle‐plastic transition conditions. The occurrence of laboratory earthquakes associated with increasing plastic deformation is promoted at high confining pressure, on smooth fault interfaces, or when the loading rate is slow. In a rather counterintuitive manner, increasing initial roughness promotes stable sliding and a larger amount of plastic deformation. Furthermore, we show that stable sliding tends to produce mirror‐like surfaces, while stick‐slips are associated with matte surfaces, on which the size of the asperities grows with increasing confining pressure. Finally, our results seem to reveal the influence of asperity hardness and melt viscosity on fault weakening.

Effects of sample preparation on the microstructural signatures of faulting in clay-bearing fault gouge

Clay-rich fault gouge in the principal slip zones of faults stores abundant water within pores and between clay interlayers. During the preparation of thin-sections and rock chips for microstructural observations, fault-gouge samples are commonly air-dried at room temperature or in an oven. However, during the drying process, remnant liquid water between gouge grains produces an inter-particle adhesion force (liquid-bridge force), which rearranges the grain-to-grain structure, likely resulting in a disturbance of the original fabric. For this study, we prepared gouge samples from the Itozawa fault, northeastern Japan, using a t-butyl alcohol freeze-drying method that mitigates drying-induced fabric disturbance. We then compared the microstructure of our samples with those prepared using the conventional air-drying method. The freeze-dried samples preserve a smooth fault plane, clearly defined nanoparticles, and well-developed shear-sense indicators, including slickenlines and Riedel shear planes. In contrast, the air-dried samples underwent shrinkage during drying, which distorted the geometry of the fault plane. These air-dried samples lack nanoparticles and display only a weak shear fabric. We conclude that microstructural observations on samples prepared using the t-butyl alcohol freeze-drying method, compared with conventional air-drying, could preserve more evidence for the retrieval of fault information, including the kinematics, slip stability, and dynamic weakening mechanism of a fault.

Sliding mode observer for fault reconstruction of time-delay and sampled-output systems – a Time Shift Approach

This paper presents a novel Time Shift Approach for actuator fault reconstruction of systems with output time-delay based on a sliding mode observer (SMO). Arbitrary output delay duration, possibly uncertain and time-varying, can be considered. Ideal sliding mode is made possible even in the presence of delay. This is particularly important since chattering problems can be circumvented so that the proposed approach is conveniently applicable to the case of sampled-output systems. In order to allow a smooth fault reconstruction using noncausal filters, an intentional output delay of one or more sampling periods is introduced. Numerical simulations illustrate the effectiveness of the proposed methodology.

Coseismic Radiation of the 2008 Mw 7.9 Wenchuan Earthquake and Its Relationship to Fault Complexities

The long rupture length and complex fault geometries of the 2008 Mw 7.9 Wenchuan earthquake offer a great opportunity to investigate the relationship between coseismic radiation and fault complexities. In this study, we first improved the compressive-sensing back-projection method (CS-BP) with window time adjustment and parallel computation to enhance its spatial–temporal resolution and computational efficiency. Then we applied the CS-BP to invert for the coseismic radiation of the Wenchuan earthquake in multiple frequency bands. Our results show that the coseismic radiation of the Wenchuan earthquake is frequency-dependent and strongly associated with the fault geometries: low-frequency radiation sources are concentrated on smooth fault segments with large coseismic slip, while higher-frequency sources are mainly related to fault complexities (e.g., Xiaoyudong fault, Gaochuan fault-bend, Leigu fault-bend, and Nanba fault step-over). An acceleration of rupture speed from ~ 1.5 to ~ 3.0 km/s is also observed the near Xiaoyudong fault, where strong high-frequency energies were radiated. Finally, we discuss the role of rupture speed in mediating the relationship between coseismic radiation and fault complexities. On one hand, sudden changes of rupture speeds contribute to strong high-frequency radiation. On the other hand, fault complexities can also strongly affect the rupture speed: a sudden break of a large asperity is able to propel the acceleration, while rupture front may decelerate when it jumps across geometric barriers.

The Slip Behavior and Source Parameters for Spontaneous Slip Events on Rough Faults Subjected to Slow Tectonic Loading

AbstractWe study the response to slow tectonic loading of rough faults governed by velocity weakening rate and state friction, using a 2‐D plane strain model. Our numerical approach accounts for all stages in the seismic cycle, and in each simulation we model a sequence of two earthquakes or more. We focus on the global behavior of the faults and find that as the roughness amplitude, br, increases and the minimum wavelength of roughness decreases, there is a transition from seismic slip to aseismic slip, in which the load on the fault is released by more slip events but with lower slip rate, lower seismic moment per unit length, M0,1d, and lower average static stress drop on the fault, Δτt. Even larger decreases with roughness are observed when these source parameters are estimated only for the dynamic stage of the rupture. For br ≤ 0.002, the source parameters M0,1d and Δτt decrease mutually and the relationship between Δτt and the average fault strain is similar to that of a smooth fault. For faults with larger values of br that are completely ruptured during the slip events, the average fault strain generally decreases more rapidly with roughness than Δτt.

Rupture preparation process controlled by surface roughness on meter-scale laboratory fault

We investigate the effect of fault surface roughness on rupture preparation characteristics using meter-scale metagabbro specimens. We repeatedly conducted the experiments with the same pair of rock specimens to make the fault surface rough. We obtained three experimental results under the same experimental conditions (6.7 MPa of normal stress and 0.01 mm/s of loading rate) but at different roughness conditions (smooth, moderately roughened, and heavily roughened). During each experiment, we observed many stick-slip events preceded by precursory slow slip. We investigated when and where slow slip initiated by using the strain gauge data processed by the Kalman filter algorithm. The observed rupture preparation processes on the smooth fault (i.e. the first experiment among the three) showed high repeatability of the spatiotemporal distributions of slow slip initiation. Local stress measurements revealed that slow slip initiated around the region where the ratio of shear to normal stress (τ/σ) was the highest as expected from finite element method (FEM) modeling. However, the exact location of slow slip initiation was where τ/σ became locally minimum, probably due to the frictional heterogeneity. In the experiment on the moderately roughened fault, some irregular events were observed, though the basic characteristics of other regular events were similar to those on the smooth fault. Local stress data revealed that the spatiotemporal characteristics of slow slip initiation and the resulting τ/σ drop for irregular events were different from those for regular ones even under similar stress conditions. On the heavily roughened fault, the location of slow slip initiation was not consistent with τ/σ anymore because of the highly heterogeneous static friction on the fault, which also decreased the repeatability of spatiotemporal distributions of slow slip initiation. These results suggest that fault surface roughness strongly controls the rupture preparation process, and generally increases its complexity with the degree of roughness.

Control of fault plane geometry on the formation of a normal fault-related anticline: an experimental approach

In one of the largest oil-gas fields in Daqing, China, the anticlines are important structures that hold natural gas. The origin of the symmetric anticlines, which have bends on both the limbs, remains under debate. This is especially true in the case of the anticline in Xujiaweizi (XJWZ), which has recently been the focus of gas exploration. A compressive force introduced by a ramp/flat fault was suggested as its origin of formation; however, this is inconsistent with the reconstruction of the regional stress fields, which show an extensive environment. An alternative explanation suggests a normal fault-related fold under extensive stress. However, this mechanism has difficulty explaining the very localized, rather than wide-spread, development of the anticline along the proposed controlling normal fault. The well-developed bends on both limbs of the anticline are also very different from the typical roll-over anticline. Here, we conduct an experimental study showing that the very localized development of the bent-on-both-limbs anticline is controlled by the geometry of the underlying fault-plane. A ramp/flat fault plane can introduce an anticline with bends on both limbs, while a smooth fault plane will develop a roll-over anticline with a bend on only one limb.

Dehydration of lawsonite could directly trigger earthquakes in subducting oceanic crust.

Intermediate-depth earthquakes in cold subduction zones are observed within the subducting oceanic crust, as well as the mantle. In contrast, intermediate-depth earthquakes in hot subduction zones predominantly occur just below the Mohorovičić discontinuity. These observations have stimulated interest in relationships between blueschist-facies metamorphism and seismicity, particularly through dehydration reactions involving the mineral lawsonite. Here we conducted deformation experiments on lawsonite, while monitoring acoustic emissions, in a Griggs-type deformation apparatus. The temperature was increased above the thermal stability of lawsonite, while the sample was deforming, to test whether the lawsonite dehydration reaction induces unstable fault slip. In contrast to similar tests on antigorite, unstable fault slip (that is, stick-slip) occurred during dehydration reactions in the lawsonite and acoustic emission signals were continuously observed. Microstructural observations indicate that strain is highly localized along the fault (R1 and B shears), and that the fault surface develops slickensides (very smooth fault surfaces polished by frictional sliding). The unloading slope during the unstable slip follows the stiffness of the apparatus at all experimental conditions, regardless of the strain rate and temperature ramping rate. A thermomechanical scaling factor for the experiments is within the range estimated for natural subduction zones, indicating the potential for unstable frictional sliding within natural lawsonite layers.

Rupture complexity and the supershear transition on rough faults

AbstractField investigations suggest that supershear earthquakes occur on geometrically simple, smooth fault segments. In contrast, dynamic rupture simulations show how heterogeneity of stress, strength, and fault geometry can trigger supershear transitions, as well as other complex rupture styles. Here we examine the Fang and Dunham (2013) ensemble of 2‐D plane strain dynamic ruptures on fractally rough faults subject to strongly rate weakening friction laws to document the effect of fault roughness and prestress on rupture behavior. Roughness gives rise to extremely diverse rupture styles, such as rupture arrests, secondary slip pulses that rerupture previously slipped fault sections, and supershear transitions. Even when the prestress is below the Burridge‐Andrews threshold for supershear on planar faults with uniform stress and strength conditions, supershear transitions are observed. A statistical analysis of the rupture velocity distribution reveals that supershear transients become increasingly likely at higher stress levels and on rougher faults. We examine individual ruptures and identify recurrent patterns for the supershear transition. While some transitions occur on fault segments that are favorably oriented in the background stress field, other transitions happen at the initiation of or after propagation through an unfavorable bend. We conclude that supershear transients are indeed favored by geometric complexity. In contrast, sustained supershear propagation is most common on segments that are locally smoother than average. Because rupture style is so sensitive to both background stress and small‐scale details of the fault geometry, it seems unlikely that field maps of fault traces will provide reliable deterministic predictions of supershear propagation on specific fault segments.

Gravity anomalies, crustal structure, and seismicity at subduction zones: 1. Seafloor roughness and subducting relief

AbstractAn ensemble averaging technique is used to remove the long‐wavelength topography and gravity field from subduction zones. >200 residual bathymetric and gravimetric anomalies are interpreted within fore arcs, many of which are attributed to the tectonic structure of the subducting plate. The residual‐gravimetric expression of subducting fracture zones extends >200 km landward of the trench axis. The bathymetric expression of subducting seamounts with height ≥1 km and area ≥500 km2 (N=36), and aseismic ridges (N>10), is largest near the trench (within 70 km) and above shallow subducting slab depths (SLAB1.0 <17 km). Subducting seamounts are similar in wavelength, amplitude, and morphology to unsubducted seamounts. Morphology, spatial distributions, and reduced levels of seismicity are considered inconsistent with mechanical models proposing wholesale decapitation, and the association of subducting seamounts with large‐earthquakes. Subducting aseismic ridges are associated with uplift and steepening of the outer fore arc, a gradual reduction in residual bathymetric expression across the inner fore arc, and a local increase in the width and elevation of the volcanic‐arc/orogen. These contrasting expressions reflect the influence of margin‐normal variations in rigidity on where and how the upper plate deforms, both to accommodate subducting relief and in response to stresses transmitted across the plate interface. The outer fore arc and arc have lower rigidity due to fracturing and thermal weakening, respectively. Similar associations with complex earthquakes and fault creep suggest aseismic ridge subduction may also be accommodated by the development and evolution of a broad fracture network, the geometrical strength of which may exceed the locking strength of a smooth fault.

Frictional strength and wear-rate of carbonate faults during high-velocity, steady-state sliding

We ran an extensive series of shear experiments to test the effect of shear velocity and normal stress on wear-rate and frictional strength. The experiments were conducted on three types of carbonate samples with a rotary shear apparatus on solid, ring-shaped rock samples that slipped for displacements up to tens of meters at slip velocity of V=0.002–0.96 m/s, and normal stress σn=0.25–6.9 MPa.The analysis reveals that during steady-state stage, the values of wear-rate and frictional strength depend on both slip velocity and normal stress. The wear-rates at low slip velocity show linear relations to the normal stress (Archardʼs model), however, at high velocity, V>0.5 m/s, the wear-rates are independent of the normal stress, and may vanish at the highest velocity and normal stress of the present experiments. The steady-state friction coefficient, μ, correlates best with the experimental power-density (= shear stress ⋅ slip velocity). We recognized three friction regimes: high μ>0.8 at low power-density, low μ∼0.3 at high power-density, and a transition regime of fast drop of friction coefficient as the power-density increases from 0.03 to 0.3 MW/m2. Runs of low power-density (high friction) displayed fault surfaces covered with fine-grained gouge, whereas runs of high power-density (low friction) displayed shiny, smooth fault surfaces. We interpret the observed intensity variations of wear-rate and frictional strength as indicating a brittle to ductile transition associated with frictional heating.

Fault Signals Drawn from Strong Interference Signals in Mineral Well Based on Least-Square Wavelet Transform

A novel wavelet transform — the least-square transform is proposed in the paper, which is used to extract weak fault signals from strong interference signals in coal mineral well. With the help of the least-square transform, a smooth fault signal curve can be got from the fault signal with high noise. Strong interference signals in the coal mineral well give a so serious problem on the work of general electric investigation that sometime process of observing useful fault signal cannot be finished. For extracting the weak effective fault signals from strong interferences signals in coal mineral well, original measured signals are decomposed by the least-square transform wavelet transform proposed in the paper, and the weak effective fault signals are extracted successfully. Experimental results show that the approach method is feasible and efficient.

Probabilistic prediction of rupture length, slip and seismic ground motions for an ongoing rupture: implications for early warning for large earthquakes

Earthquake Early Warning (EEW) predicts future ground shaking based on presently available data. Long ruptures present the best opportunities for EEW since many heavily shaken areas are distant from the earthquake epicentre and may receive long warning times. Predicting the shaking from large earthquakes, however, requires some estimate of the likelihood of the future evolution of an ongoing rupture. An EEW system that anticipates future rupture using the present magnitude (or rupture length) together with the Gutenberg-Richter frequency-size statistics will likely never predict a large earthquake, because of the rare occurrence of ‘extreme events’. However, it seems reasonable to assume that large slip amplitudes increase the probability for evolving into a large earthquake. To investigate the relationship between the slip and the eventual size of an ongoing rupture, we simulate suites of 1-D rupture series from stochastic models of spatially heterogeneous slip. We find that while large slip amplitudes increase the probability for the continuation of a rupture and the possible evolution into a ‘Big One’, the recognition that rupture is occurring on a spatially smooth fault has an even stronger effect. We conclude that an EEW system for large earthquakes needs some mechanism for the rapid recognition of the causative fault (e.g., from real-time GPS measurements) and consideration of its ‘smoothness’. An EEW system for large earthquakes on smooth faults, such as the San Andreas Fault, could be implemented in two ways: the system could issue a warning, whenever slip on the fault exceeds a few metres, because the probability for a large earthquake is high and strong shaking is expected to occur in large areas around the fault. A more sophisticated EEW system could use the present slip on the fault to estimate the future slip evolution and final rupture dimensions, and (using this information) could provide probabilistic predictions of seismic ground motions along the evolving rupture. The decision on whether an EEW system should be realized in the first or in the second way (or in a combination of both) is user-specific.