The response of creeping parts of the San Andreas fault to earthquakes on nearby faults: Two examples

Rates of shallow slip on creeping sections of the San Andreas fault have been perturbed on a number of occasions by earthquakes occurring on nearby faults. One example of such perturbations occurred during the 26 January 1986 magnitude 5.3 Tres Pinos earthquake located about 10 km southeast of Hollister, California. Seven creepmeters on the San Andreas fault showed creep steps either during or soon after the shock. Both left-lateral (LL) and right-lateral (RL) steps were observed. A rectangular dislocation in an elastic half-space was used to model the coseismic fault offset at the hypocenter. For a model based on the preliminary focal mechanism, the predicted changes in static shear stress on the plane of the San Andreas fault agreed in sense (LL or RL) with the observed slip directions at all seven meters; for a model based on a refined focal mechanism, six of the seven meters showed the correct sense of motion. Two possible explanations for such coseismic and postseismic steps are (1) that slip was triggered by the earthquake shaking or (2) that slip occurred in response to the changes in static stress fields accompanying the earthquake. In the Tres Pinos example, the observed steps may have been of both the triggered and responsive kinds. A second example is provided by the 2 May 1983 magnitude 6.7 Coalinga earthquake, which profoundly altered slip rates at five creepmeters on the San Andreas fault for a period of months to years. The XMM1 meter 9 km northwest of Parkfield, California recorded LL creep for more than a year after the event. To simulate the temporal behavior of the XMM1 meter and to view the stress perturbation provided by the Coalinga earthquake in the context of steady-state deformation on the San Andreas fault, a simple time-evolving dislocation model was constructed. The model was driven by a single long vertical dislocation below 15 km in depth, that was forced to slip at 35 mm/yr in a RL sense. A dislocation element placed in the seismogenic layer under XMM1 was given a finite breaking strength of sufficient magnitude to produce a Parkfield-like earthquake every 22 years. When stress changes equivalent to a Coalinga earthquake were superposed on the model running in a steady state mode, the effect was to make a segment under XMM1, that could slip in a linear viscous fashion, creep LL and to delay the onset of the next Parkfield-like earthquake by a year or more. If static stress changes imposed by earthquakes off the San Andreas can indeed advance or delay earthquakes on the San Andreas by months or years, then such changes must be considered in intermediate-term prediction efforts.     

Distribution of postseismic slip on the Calaveras fault,California, following the 1984 M6.2 Morgan Hill earthquake

Repeating earthquakes (REs) are sequences of events that have virtually identical waveforms and are interpreted to represent fault asperities driven to failure by loading from aseismic creep on the surrounding fault surface at depth. To investigate the postseismic deformation after the 1984 M6.2 Morgan Hill earthquake, we identify RE sequences occurring on the central Calaveras fault between 1984 and 2005 using a combination of cross-correlation and spectral coherence techniques. Both the accelerated slip transients due to the earthquake as well as the return to interseismic background creep rates can be imaged from our dataset. A comparison between the regions of the fault that ruptured coseismically and the locations of the REs show that REs preferentially occur in areas adjacent to the coseismic rupture. Using calculated RE-derived subsurface slip distributions at 6 months and 18 months after the mainshock, we predict surface electronic distance meter (EDM) line length changes between stations near the Morgan Hill rupture area. The RE-derived slip model underpredicts a subset of the observed line-length changes. Inclusion of transient aseismic slip below the seismogenic zone is needed to better match the measured surface deformation.     

Assessment of Creep Events as Potential Earthquake Precursors: Application to the Creeping Section of the San Andreas Fault, California

—We report the analysis of over 16 years of fault creep and seismicity data from part of the creeping section of the San Andreas fault to examine and assess the temporal association between creep events and subsequent earthquakes. The goal is to make a long-term evaluation of creep events as a potential earthquake precursor. We constructed a catalog of creep events from available digital creepmeter data and compared it to a declustered seismicity catalog for the area between San Juan Bautista and San Benito, California, for 1980 to 1996. For magnitude thresholds of 3.8 and above and time windows of 5 to 10 days, we find relatively high success rates (40% to 55% 'hits') but also very high false alarm rates (generally above 90%). These success rates are statistically significant (0.0007 < P < 0.04). We also tested the actual creep event catalog against two different types of synthetic seismicity catalogs, and found that creep events are followed closely in time by earthquakes from the real catalog far more frequently than the average for the synthetic catalogs, generally by more than two standard deviations. We find no identifiable spatial pattern between the creep events and earthquakes that are hit or missed. We conclude that there is a significant temporal correlation between creep events and subsequent small to moderate earthquakes, however that additional information (such as from other potential precursory phenomena) is required to reduce the false alarm rate to an acceptable level.     

Source parameters of central California earthquakes on February 24 and September 4, 1972

The directivity function method is combined with an earthquake sequence study to obtain the reliable estimates of rupture area and rupture velocity of two right-lateral strike-slip earthquakes occurring along the San Andreas fault zone in central California. By utilizing a significant difference in velocity structure on both sides of the main fault, a modified version of the directivity function is formulated and applied to near-field SH waves. On assuming aftershock areas following their step-wise increase with time as rupture areas, an extensive systematic search for the best fit between the observed and theoretical directivity functions is made for a combination of both source parameters. The rupture at the February 24, 1972 earthquake (M_L = 5.0) propagated unilaterally with an average velocity of about 2.3 km/sec. It produced a rectangular area, the horizontal and vertical lengths being about 5 and 2 km, respectively. The rupture at the September 4, 1972 earthquake (M_L = 4.6) was of bilateral, yielding a nearly square rupture area of which side length is about 2 km. The rupture velocity of this earthquake, though some ambiguities resulting from a lower quality of the observed directivity function, is estimated at 1.9 km/sec or less. A difference in average rupture velocity between both earthquakes might imply its dependence on the ambient tectonic environment such as represented by local stress and humidity. By taking into account a post-earthquake creep increase of about 3.0–3.5 cm observed at a creepmeter station situated over the focus of the February 24, 1972 earthquake, its stress drop and seismic moment are estimated at about 10 bar and 10²³ dyn · cm, respectively. The above procedure has a broad applicability for recovering reliable estimates of source parameters, especially when it is combined with a synthetic approach.     

1966年邢台强震群震前、 同震和震后断层运动特征分析

在1966年3月邢台强震群水准观测基础上, 本文首先对同震位移和数值模拟结果进行了对比, 认为邢台强震群发震断层浅层是牛家桥阳断层和永福庄断层, 深部发震断层是倾向北西的北东向断层; 然后以区域水准观测为约束, 通过反演给出了邢台强震群震前断层无震滑移、 同震断层位错和震后断层余滑空间分布。 结果表明, 震前显著无震滑移主要发生在3月8日6.8级前震的发震断层上, 而3月20日宁晋7.2级和6.7级地震发震断层深部同震位错相对弱的断层段, 震前也存在较为明显的无震滑移; 震后断层余滑主要分布在束鹿凹陷西侧的牛家桥阳断层。 从邢台强震群过程中断层运动时空演化特征可以看出, 强震破裂成核过程中震前断层无震滑移既可能发生于前震发震断层, 也可能发生于主震发震断层, 震后断层位移空间分布与同震位错具有互补性; 而震前断层无震滑移是否为有关震前前兆的真正原因, 是值得进一步深入研究的问题。     

Application of the stress evolutionary model along the Xiaojiang fault zone in Yunnan Province,Southeast China

The Xiaojiang fault zone constitutes part of the major Xianshuihe-Xiaojiang left lateral structure that bounds the rhombic-shaped block of Yunnan-Sichuan to the east. Long strike slip fault zones that have repeatedly accommodated intense seismic activity, constitute a basic feature of southeast China. Known historical earthquakes to have struck the study area are the 1713 Xundian of M6.8, 1725 Wanshou mountain of M6.8, the 1733 Dongchuan of M7.8, and the strongest one, the 1833 Songming of M8.0. Although instrumental record did not report events of this magnitude class, the 18th century clustering as well as the 19th century large event prompted the investigation of stress transfer along this fault zone. Coulomb stress changes were calculated assuming that earthquakes can be modeled as static dislocations in an elastic half-space, and taking into account both the coseismic slip in strong (M ≥ 6.8) earthquakes and the slow tectonic stress buildup along the major fault segments. Geological and geodetic data are used to infer the geometry of these faults and long term slip rates on them, as well as for the fault segments that slipped. Evidence is presented that the strong historical events as well as the ones of smaller magnitude that occurred during the instrumental era, are located in areas where the static stress was enhanced. By extending the calculations up to present, possible sites for future strong events are identified.     

A viscoelastic relaxation model for post-seismic deformation from the San Francisco earthquake of 1906

We present a realistic model of the San Andreas fault zone. We propose that aseismic ground displacement is a sum of visco-elastic relaxation following large earthquakes, transient fault slip, steady fault slip and a large-scale relative plate motion. We used the model to explain the aseismic ground displacements observed after the San Francisco earthquake of 1906.The data do not resolve the question of which is the dominant mechanism, but viscoelastic relaxation can contribute a significant fraction of the displacement if the elastic plate thickness is 50 km or less. If the relative plate motion is taken to be 5.5 cm/yr, as found from plate rotation pole studies, then the zone of significant shearing in the mantle is probably at least 100 km thick beneath California.     

基于形变观测分析2011年日本9.0级地震与断层运动间关系

2011年3月11日日本发生9.0级地震,本文以此次地震的震间、同震和震后形变观测为约束,依据不同时段断层运动空间分布特征分析日本海沟地区强震与断层运动间关系.震间日本海沟地区,断层运动闭锁线深度约为60km,闭锁线以上从深到浅依次为断层运动强闭锁段、无震滑移段和弱闭锁段.由同震位错反演结果,2011年日本9.0级地震同震存在深浅两个滑移极值区,同震较浅的滑移极值区(同震位错量10~50m,深度小于30km)震间为断层弱闭锁段;同震较深的滑移极值区(同震位错量10~20m,深度在40km左右)震间为断层强闭锁段;而在两者之间的过渡带同震位错相对较小,震间断层运动表现为无震滑移.震后初期断层运动主要分布在在闭锁线以上的同震较深滑移极值区,而同震较浅的滑移极值区能量释放比较彻底,断层震后余滑量相对较小.依据本文同震和震间断层运动反演结果,震间强闭锁段积累10m同震位错需要100多年时间,与该区域历史上7级地震活动复发周期相当;震间弱闭锁段积累30~50m同震位错约需要300~600年时间,与相关研究给出的日本海沟9级左右地震复发周期比较一致.在实际孕震能力判定的工作中,由于不同性质的断层段在同震过程中会表现更多的组合形式,断层发震能力判定结果存在更多的不确定性,但利用区域形变观测等资料给出震间断层运动特征的研究工作对于断层强震发震能力的判定具有非常重要的实际意义.     

Clay gouges in the San Andreas Fault System and their possible implications

Summary In the northern and central sections of the San Andreas Fault Zone, and along Calaveras and Hayward faults, clay gouges have been found to occur on the surface and at shallow depths.It is consistent with the available geochemical data that such gouges can exist at depths down to 10 km. If extensive gouge materials exist in a fault zone then their properties will determine, to a large extent, the behavior of the fault. From known properties of clays in the presence of water we can infer that, in such cases, the tectonic stress and the stress drops for earthquakes will be low and substantial creep will take place before earthquakes.     

Spatio-temporal Slip, and Stress Level on the Faults within the Western Foothills of Taiwan: Implications for Fault Frictional Properties

We use preseismic, coseismic, and postseismic GPS data of the 1999 Chi-Chi earthquake to infer spatio-temporal variation of fault slip and frictional behavior on the Chelungpu fault. The geodetic data shows that coseismic slip during the Chi-Chi earthquake occurred within a patch that was locked in the period preceding the earthquake, and that afterslip occurred dominantly downdip from the ruptured area. To first-order, the observed pattern and the temporal evolution of afterslip is consistent with models of the seismic cycle based on rate-and-state friction. Comparison with the distribution of temperature on the fault derived from thermo-kinematic modeling shows that aseismic slip becomes dominant where temperature is estimated to exceed 200° at depth. This inference is consistent with the temperature induced transition from velocity-weakening to velocity-strengthening friction that is observed in laboratory experiments on quartzo-feldspathic rocks. The time evolution of afterslip is consistent with afterslip being governed by velocity-strengthening frictional sliding. The dependency of friction, μ, on the sliding velocity, V, is estimated to be ${{\partial \mu }/{\partial \, {\rm ln}\, V}} = 8 \times 10^{ - 3}$ . We report an azimuthal difference of about 10–20° between preseismic and postseismic GPS velocities, which we interpret to reflect the very low shear stress on the creeping portion of the décollement beneath the Central Range, of the order of 1–3 MPa, implying a very low friction of about 0.01. This study highlights the importance of temperature and pore pressure in determining fault frictional sliding.     

First study of fault trench stratigraphy at Mt. Etna volcano, Southern Italy: understanding Holocene surface faulting along the Moscarello fault

Paleoseismology, the study of past earthquakes based on their geological record in the stratigraphy and landscape, is a successful newly developing field of research. The application of fault trench studies in volcanic environments is one of the youngest branches of paleoseismology. In this paper, we present the results of the first exploratory trenches excavated at Mt. Etna in Sicily, the largest European volcano. Modern surface faulting at Etna is a very well known feature, which poses significant hazard to the local community, both in terms of ground displacement of essential lifelines and ground shacking from frequent damaging earthquakes. However, while the geomorphology and the seismicity of the active fault in the Etna region consistently show very high rates of tectonic activity, the Holocene cumulative throw and slip-rates, along with the nature (coseismic vs. creeping fault slip), dimension and timing of the displacement events, are still poorly constrained. For this purpose, we selected as a sample area the Moscarello fault, one of the most outstanding segment of the Timpe system of active normal faults in the volcano’s lower eastern flank. Displaced landforms and volcanic units at the Fondo Macchia basin, in the central sector of this fault, indicate some hundreds of meters of vertical offset in the last ca. 80 kyr, with a long-term slip-rate substantially higher than 1.5–2.0 mm/yr. According to the historical sources and instrumental observations, the Moscarello fault ruptured four times in the last 150 years during shallow (H < 5 km) and moderate magnitude (M < 4.8) earthquakes. These events were associated with severe damage in a narrow epicentral area (macroseismic intensities up to the IX–X grade of the MSK scale) and extensive surface faulting (end-to-end rupture length up to 6 km, vertical offsets up to 90 cm). This clearly indicates very high modern rates of deformation along this fault. We conducted trench investigations at the Fondo Macchia site, in a point where eyewitnesses observed ca. 20 cm of coseismic vertical displacement after the April 21, 1971, M_s=3.7, earthquake. The excavated sections provided direct stratigraphic evidence for a vertical slip-rate of 1.4–2.7 mm/yr in the last ca. 6 kyr. This should be regarded as a minimum slip-rate for the central section of the fault. We explored a single scarp at a single site, while we know from recent historical observations that several parallel scarps may rupture coseismically at Fondo Macchia. Thus, the relevant deformation rate documented for the modern period might be likely extended back in the past to a time-span of some thousands of years at least. As expected, for such a volcanic environment, the activity rates of the Moscarello fault are also significantly higher than for the Apennines normal faults, typically showing slip-rates lower than 1 mm/yr. The agriculturally reworked trench hangingwall stratigraphy did not allow to recognize individual displacement events. Nevertheless, the sedimentary structures observed in the trench footwall strongly suggest that, as for the last 150–200 years of detailed historical record, fault behavior at Fondo Macchia is governed by coseismic surface displacement rather than fault creep. This research confirms that paleoseismology techniques can be effectively applied also in active volcanic environments, typically characterized by rheology and, consequently, seismicity and fault dynamics very different from those of other tectonic environments in which paleoseismology has been firstly developed and is today extensively applied.     

Instability model for recurring large and great earthquakes in southern California

The locked section of the San Andreas fault in southern California has experienced a number of large and great earthquakes in the past, and thus is expected to have more in the future. To estimate the location, time, and slip of the next few earthquakes, an earthquake instability model is formulated. The model is similar to one recently developed for moderate earthquakes on the San Andreas fault near Parkfield, California. In both models, unstable faulting (the earthquake analog) is caused by failure of all or part of a patch of brittle, strain-softening fault zone. In the present model the patch extends downward from the ground surface to about 12 km depth, and extends 500 km along strike from Parkfield to the Salton Sea. The variation of patch strength along strike is adjusted by trial until the computed sequence of instabilities matches the sequence of large and great earthquakes sincea.d. 1080 reported by Sieh and others. The last earthquake was theM=8.3 Ft. Tejon event in 1857. The resulting strength variation has five contiguous sections of alternately low and high strength. From north to south, the approximate locations of the sections are: (1) Parkfield to Bitterwater Valley, (2) Bitterwater Valley to Lake Hughes, (3) Lake Hughes to San Bernardino, (4) San Bernardino to Palm Springs, and (5) Palm Springs to the Salton Sea. Sections 1, 3, and 5 have strengths between 53 and 88 bars; sections 2 and 4 have strengths between 164 and 193 bars. Patch section ends and unstable rupture ends usually coincide, although one or more adjacent patch sections may fail unstably at once. The model predicts that the next sections of the fault to slip unstably will be 1, 3, and 5; the order and dates depend on the assumed length of an earthquake rupture in about 1700.     

Coseismic deformation of the 2021 MW7.4 Maduo earthquake from joint inversion of InSAR,GPS, and teleseismic data

The M_W7.4 Maduo earthquake occurred on 22 May 2021 at 02:04 CST with a large-expansion surface rupture. This earthquake was located in the Bayan Har block at the eastern Tibetan Plateau, where eight earthquakes of M_S >7.0 have occurred in the past 25 years. Here, we combined interferometric synthetic aperture radar, GPS, and teleseismic data to study the coseismic slip distribution, fault geometry, and dynamic source rupture process of the Maduo earthquake. We found that the overall coseismic deformation field of the Maduo earthquake is distributed in the NWW-SEE direction along 285°. There was slight bending at the western end and two branches at the eastern end. The maximum slip is located near the eastern bending area on the northern branch of the fault system. The rupture nucleated on the Jiangcuo fault and propagated approximately 160 km along-strike in both the NWW and SEE directions. The characteristic source rupture process of the Maduo earthquake is similar to that of the 2010 M_W6.8 Yushu earthquake, indicating that similar earthquakes with large-expansion surface ruptures and small shallow slip deficits can occur on both the internal fault and boundary fault of the Bayan Har block.     

Possible strain events reflected in water levels in wells along San Jacinto fault zone,southern California

Water levels have been monitored in wells along the San Jacinto fault zone since 1977. The three largest earthquakes to occur within 30 km of the segment of the San Jacinto fault zone being monitored with continuous recorders showed magnitudesM of 4.5, 4.8, and 5.5. Two wells in Borrego Valley, 31 to 32 km southeast of theM=5.5 earthquake on 25 February 1980, showed anomalous spikes recording a probable strain event 88 hours before the earthquake. Two other wells 12 km northwest of the epicenter showed no water-level anomalies. No water-level anomalies preceded theM=4.8 earthquake near Anza on 15 June 1982. Anomalous water-level fluctuations occurred in a well near Ocotillo Wells, 13 km northeast of theM=4.5 earthquake on 22 March 1982, 19 to 23 days prior to the earthquake. Similar fluctuations in other wells have not been followed by sizable earthquakes. A simultaneous drop in water level occurred in four wells on 8 September 1982; this possible strain event was not associated with a sizable earthquake. The presumed strain events occur only in wells that show earth tides and may have been the result of creep on strands of the San Jacinto fault zone. Although water-level anomalies have occurred in only one or two wells prior to two out of three moderate (M=4.5–5.5) earthquakes, the simultaneous drop in water level on 8 September 1982 and the spikes in two wells before theM=5.5 earthquake on 25 February 1980 suggest that wells responsive to earth tides may detect strain events.     

Stress accumulation and release on the San Andreas fault

Summary The San Andreas fault can be divided into locked and free sections. On the locked sections accumulated slip is released in great earthquakes. On the free sections slip is occurring continuously either aseismically or during smaller earthquakes. Stress drops during earthquakes can be estimated from the ratio of short to long period amplitudes and from surface strain. Surface heat flow may provide an upper bound on the absolute stress. The failure or yield stress must reach a maximum at some depth on the fault. This maximum may occur in the near-surface brittle zone or deeper in the plastic zone of the fault. The historic distribution of seismic activity provides information on the stress level. The accumulation of strain and stress on the fault can be predicted using elastic theory. It is necessary, however, to include the viscous coupling of the lithosphere to the asthenosphere in order to fully model the problem.     

InSAR数据约束的2021年5月21日云南漾濞MS6.4地震发震构造研究

运用Sentinel-1A卫星数据和D-InSAR技术,获取2021-05-21云南漾濞M_S6.4地震的同震形变场。结果显示,漾濞地震同震形变场长轴近NW展布升降轨形变场符号相反,视线向最大沉降量和抬升量为0.1 m。InSAR同震形变场反演的滑动分布主要集中在沿走向2～12 km,倾向1～9 km的范围内,最大滑动量0.35 m,发震断层长9.8 km、宽4 km,滑动量主要集中在地下3～6 km范围内,滑动角-146.7°。同震位移场及滑动分布模型反映本次地震为发震断层的右旋走滑事件,地震破裂未达到地表。断层模型反演结果显示,矩震级为M_W6.1,发震断层以北西走向右旋走滑运动为主,初步认为本次M_W6.1地震发震断裂可能是一条NW向的维西—乔后断裂西侧的隐伏次生断裂。     

Constraints from precariously balanced rocks on preferred rupture directions for large earthquakes on the southern San Andreas Fault

We have compared near-fault ground motions from TeraShake simulations of M_w7.7 earthquake scenarios on the southern San Andreas Fault with precariously balanced rock locations. The TeraShake scenarios with different directions of rupture generate radically different ground motions to the northwest of the Los Angeles Basin, primarily because of directivity effects, and thus provide constraints on the ground motion and rupture direction for the latest (1690) large event on that section of the San Andreas Fault. Due to the large directional near-field ground motions predicted by the simulations, we expect the precariously balanced rocks to be located primarily in the backward rupture direction or near the epicenter. Preliminary results favor persistent nucleation at or slightly northwest of the San Gorgonia Pass fault zone for large earthquakes on the southern San Andreas Fault.     

Coseismic stress parameter of three California Earthquakes derived from the stochastic finite fault technique

Stochastic finite fault modeling is used to derive the coseismic stress parameter distribution on the fault surface of three well-recorded California earthquakes: M7.0, 1989, Loma Prieta; M7.3, 1992, Landers; and M6.7, 1994, Northridge. Classical waveform inversion techniques are inherently more powerful than stochastic modeling as a means of deriving detailed source parameters. However, the application of stochastic methods to the source modeling problem is useful to: (1) explore and calibrate the limitations and boundaries of stochastic modeling, (2) understand its relationship to more deterministically based techniques, and (3) provide a view of the source radiation not available from deterministic modeling. The stress parameter distribution for the M7.0 1989 Loma Prieta earthquake fault shows a concentration of stress in the lower part of the northwest side of the fault and another concentration in the upper southeast side of the fault, with an average stress parameter of 80 bars over the fault surface. The stress parameter distribution for the M7.3 1992 Landers earthquake fault shows a gradual increase of stress starting from the southeast side of the fault, close to the hypocenter, towards the center. The maximum stress occurs in the lower central part of the modeled fault surface. The average stress parameter is 70 bars for the Landers earthquake. The stress parameter distribution of the M6.7 1994 Northridge earthquake shows a concentration at the lower southeast end of the fault surface, extending toward the center of the fault surface and stretching to the northwest end. The average stress parameter is 80 bars for Northridge earthquake. The stress parameter distributions derived in this study by stochastic finite-fault modeling of high-frequency motions show considerable similarity to many of the slip distributions provided by different research groups for the same earthquakes, suggesting that the derivation of stress parameter distribution on a fault surface by the method applied in this study is reliable and closely tied to slip on the fault.     

Earthquakes: Recurrence and Interoccurrence Times

The purpose of this paper is to discuss the statistical distributions of recurrence times of earthquakes. Recurrence times are the time intervals between successive earthquakes at a specified location on a specified fault. Although a number of statistical distributions have been proposed for recurrence times, we argue in favor of the Weibull distribution. The Weibull distribution is the only distribution that has a scale-invariant hazard function. We consider three sets of characteristic earthquakes on the San Andreas fault: (1) The Parkfield earthquakes, (2) the sequence of earthquakes identified by paleoseismic studies at the Wrightwood site, and (3) an example of a sequence of micro-repeating earthquakes at a site near San Juan Bautista. In each case we make a comparison with the applicable Weibull distribution. The number of earthquakes in each of these sequences is too small to make definitive conclusions. To overcome this difficulty we consider a sequence of earthquakes obtained from a one million year “Virtual California” simulation of San Andreas earthquakes. Very good agreement with a Weibull distribution is found. We also obtain recurrence statistics for two other model studies. The first is a modified forest-fire model and the second is a slider-block model. In both cases good agreements with Weibull distributions are obtained. Our conclusion is that the Weibull distribution is the preferred distribution for estimating the risk of future earthquakes on the San Andreas fault and elsewhere.     

Three-dimensional P-wave Velocity Structure of the Bear Valley Region of Central California

—The three-dimensional P-wave velocity structure of the Bear Valley region of central California is determined by applying a circular ray-tracing technique to 1735 P-wave arrivals from 108 locally recorded earthquakes. Comparison of the results obtained from one-dimensional and laterally varying starting models shows that many of the features in the structure determined are fairly insensitive to the choice of the starting model. Velocities associated with the Gabilan granites southwest of the San Andreas Fault are slightly higher than those in the Franciscan formation to the northeast, and these two features are separated in the southern part of the region by a narrow fault zone with very low velocities. In the southeastern part of the region, where the Gabilan granites do not abut the San Andreas Fault, the low velocities of the fault zone cross over to the southwestern side of the fault. They also appear to extend to depths of at least 15km, thus locally reversing the contrast across the San Andreas Fault that prevails farther to the northwest. In the northwestern part of the region, the low velocities of the fault zone split and follow the surface traces of the San Andreas and Calaveras Faults, but do not appear to extend to depths much deeper than about 6km. There also appears to be a well-defined contrast in structure in the middle of the Santa Clara Valley, suggesting the existence of a fault in the basement of the valley that may be a southern extension of the Sargent Fault into this region. Relocated hypocenters beneath the San Andreas Fault cluster in a zone that dips about 80° southwest and intersects the surface trace of the fault in the southern part of region.