Note on rain-triggered earthquakes and their dependence on karst geology

Recently reported rain-triggered seismicity from three separate storms occurred exclusively in karst geology. In this paper, I discuss how the hydrogeology of karst controls rain-triggered seismicity by channeling of the watershed after intense rainfall directly into the karst network. Such channeling results in very large increases in hydraulic head, and more importantly, substantially increases the vertical stress acting on the underlying pore-elastic media. Rapid loading upon a pore-elastic media induces seismicity by increasing pore pressure at depth in a manner similar to that observed from reservoir impounding. Using a simple 1-D model of a pore-elastic medium, it is shown that the instantaneous fluid pressure increase at depth is a substantial fraction of the pressure step applied at the boundary, followed by time-dependent pore pressure increases associated with the typical linear diffusion problem. These results have implications for the change in fluid pressure necessary to trigger earthquakes, and leads to the following hypothesis to be tested: Unambiguous rain-triggered seismicity will only occur in karst regions.     

New seismogenic source and deep structures revealed by the 1999 Chia-yi earthquake sequence in southwestern Taiwan

In a tectonically active setting large earthquakes are always threats; however, they may also be useful in elucidating the subsurface geology. Instrumentally recorded seismicity is, therefore, widely utilized to extend our knowledge into the deeper crust, especially where basement is involved. It is because the earthquakes are triggered by underground stress changes that usually corresponding to the framework of geological structures. Hidden faults, therefore, can be recognized and their extension as well as orientation can be estimated. Both above are of relevance for assessment on seismic hazard of a region, since the active faults are supposed to be re-activated and cause large earthquakes. In this study, we analysed the 1999 October 22 earthquake sequence that occurred in southwestern Taiwan. Two major seismicity clusters were identified with spatial distribution between depths of 10 and 16 km. One cluster is nearly vertical and striking 032°, corresponding to the strike-slip Meishan fault (MSF) that generated the 1906 surface rupture. Another cluster strikes 190° and dips 64° to the west, which is interpreted as west-vergent reverse fault, in contrast to previous expectation of east vergence. Our analysis of the focal solutions of all the larger earthquakes in the 1999 sequence with the 3-D distribution of all the earthquakes over the period 1990–2004 allows us reinterpret the structural framework and suggest previously unreognized seismogenic sources in this area. We accordingly suggest: (1) multiple detachment faults are present in southwestern Taiwan coastal plain and (2) additional seismogenic sources consist of tear faults and backthrust faults in addition to sources associated with west-vergent fold-and-thrust belt.     

Application of the negative binomial to earthquake occurrences in the Alpide-Himalayan belt

Summary . Shallow focus earthquakes ( h ≤ 60 km) of magnitude range M = 4.0–6.0, occurred during 1954–75 in various high seismicity zones of the Alpide-Himalayan belt have been tested by the Poisson and the negative binomial laws. When the clustering of events make the simple Poisson model inapplicable in most of the high seismicity zones of the Alpide-Himalayan belt it has been shown that the negative binomial entries provide an excellent model for describing the earthquake occurrences. The chi-square ( X ²) test is employed for testing the actual observations with theoretical distributions.     

Modelling of short-interval silent slip events in deeper subduction interfaces considering the frictional properties at the unstable–stable transition regime

Recent high-resolution observations of crustal movements have revealed silent slip events (SSEs) with propagation velocities of around 10–15 km d⁻¹ and with intervals of 3–14 months along the deeper parts of the Cascadia and Nankai subduction zones. This study develops 2-D and 3-D models of these short-interval SSEs considering the frictional behaviour that was confirmed experimentally by Shimamoto for the unstable–stable transition regime. To represent this frictional behaviour, a small cut-off velocity to an evolution effect is introduced in a rate- and state-dependent friction law. When the cut-off velocity to the evolution effect is significantly smaller than that to a direct effect, steady-state friction exhibits velocity weakening at low slip velocities and velocity strengthening at high slip velocities. At the deeper Cascadia and Nankai subduction interfaces, the pore pressure is inferred to be high because of the dehydration of materials in the descending plate. Under conditions where the pore-fluid pressure is nearly equal to the lithostatic pressure and the critical weakening displacement is very small, short-interval SSEs with propagation velocities and slip velocities of 4–8 km d⁻¹ and  2 − 4 × 10⁻⁷  m s⁻¹, respectively, can be reproduced. The propagation velocity of short-interval SSEs is in proportion to the slip velocity. The results also show that during the nucleation process of large earthquakes, the occurrence of short-interval SSEs becomes irregular because of the accelerated slips that occur at the bottom of the seismogenic zone. Our results suggest that monitoring of short-interval SSEs might be useful for forecasting the main earthquakes.     

Frequency–size distribution of global seismicity seen frombroad-band radiated energy

The frequency–energy distribution of global seismicity is studied using broad-band radiated energy of shallow earthquakes from January 1987 to December 1994 estimated by NEIC. Rank-ordering statistics are applied to enhance the resolution in retrieving the power-law distribution with undersampled data, namely a few tens of events. Seen in the perspective of broad-band radiated energy with higher resolution, a single (Gutenberg–Richter-type) power-law distribution can fit the data. For earthquakes with energy larger than 10¹⁴ J, the number N of events with energy E depends on E via N∝E ^−B , with the scaling constant B = 0.64 ± 0.04, corresponding to b = 0.95 ± 0.06. This relation is different from that of scalar seismic moment, which shows a transition of power-law distributions between small and large earthquakes. To demonstrate such a difference we use the same set of earthquakes with both broad-band energy estimation and CMT estimation. It is found that for the same data set, the energy distribution and the moment distribution show different patterns. The moment distribution has a clear kink between small and large earthquakes, while the energy distribution shows a single power law with no convincing kink between small and large earthquakes. To investigate the effect of different focal mechanisms and different seismic regions, events with strike-slip mechanisms and events within the Japan–Kuril region are considered. For these subsets of events, a similar pattern exists, in which the moment distribution shows a kink between small and large earthquakes, while the energy distribution shows a single power law.     

Source parameters of large historical (1918–1962) earthquakes, South Island, New Zealand

We present the results of body waveform modelling studies for 17 earthquakes of M _w ≥5.7 occurring in the South Island, New Zealand region between 1918 and 1962, including the 1929 M _s = 7.8 Buller earthquake, the largest earthquake to have occurred in the South Island this century. These studies confirm the concept of slip partitioning in the northern South Island between strike-slip faulting in southwestern Marlborough and reverse and strike-slip faulting in the Buller region, but indicate that the zone of reverse faulting is quite localized. In the central South Island, all historical earthquakes appear to be associated with strike-slip faulting, although recent (post-1991) reverse faulting events suggest that slip partitioning also occurs within this region. The difference between historical and recent seismicity in the central South Island may also reflect stress readjustment occurring in response to the 1717 ad rupture along the Alpine fault. Within the Fiordland region (southwestern South Island) none of the historical earthquakes appears to have occurred along the Australian/Pacific plate interface, but rather they are associated with complex deformation of the subducting plate as well as with deformation of the upper (Pacific) plate. Two earthquakes in the Puysegur Bank region south of the South Island suggest that strike-slip deformation east of the Puysegur Trench is playing a major role in the tectonics of the region.     

Upper-crustal structure of the Benevento area (southern Italy): fault heterogeneities and potential for large earthquakes

The Benevento region is part of the southern Apennines seismogenic belt, which experienced large destructive seismic events both in historical and in recent times. The study area lies at the northern end of the Irpinia fault, which ruptured in 1980 with a M_s = 6.9 normal faulting event, which caused about 3000 casualties. The aims of this paper are to image lateral heterogeneities in the upper crust of the Benevento region, and to try to identify the fault segments that are expected to generate such large earthquakes. This work is motivated by the recognition that lithological heterogeneities along major fault zones, inferred from velocity anomalies, reflect the presence of fault patches that behave differently during large rupture episodes. In this paper, we define the crustal structure of the Benevento region by using the background seismicity recorded during 1991 and 1992 by a local seismic array. These data offer a unique opportunity to investigate the presence of structural discontinuities of a major seismogenic zone before the occurrence of the next large earthquake. The main result that we obtained is the delineation of two NW-trending high-velocity zones (HVZs) in the upper crust beneath the Matese limestone massif. These high velocities are interpreted as high-strength regions that extend for 30-40 km down to at least 12 km depth. The correspondence of these HVZs with the maximum intensity regions of historical earthquakes (1688 AD, 1805 AD) suggests that these anomalies delineate the extent of two fault segments of the southern Apenninic belt capable of generating M = 6.5⁻⁷ earthquakes. The lateral offset observed between the two segments from tomographic results and isoseismal areas is possibly related to transverse right-lateral faults.     

A major seismogenic fault in a 'silent area': the Castrovillari fault (southern Apennines, Italy)

Large historical earthquakes in Italy define a prominent gap in the Pollino region of the southern Apennines. Geomorphic and palaeoseismological investigations in this region show that the Castrovillari fault (CF) is a major seismogenic source that could potentially fill the southern part of this gap. The surface expression of the CF is a complex, 10–13 km long set of prominent scarps. Trenches across one scarp indicate that at least four surface-faulting earthquakes have occurred along the CF since Late Pleistocene time, each producing at least 1 m of vertical displacement. The length of the fault and the slip per event suggest M =6.5-7.0 for the palaeoearthquakes. Preliminary radiocarbon dating coupled with historical considerations imply that the most recent of these earthquakes occurred between 380 BC and 1200 AD, and probably soon after 760 AD; no evidence for this event has been found in the historical record. We estimate a minimum recurrence interval of 1170 years and a vertical slip rate of 0.2-0.5 mm yr^-1 for the CF, which indicates that the seismic behaviour of this fault is comparable to other major seismogenic faults of the central-southern Apennines. The lack of mention or the mislocation of the most recent event in the historical seismic memory of the Pollino region clearly shows that even in Italy, which has one of the longest historical records of seismicity, a seismic hazard assessment based solely on the historical record may not be completely reliable, and shows that geological investigations are critical for filling possible information gaps.     

Seismicity and seismic gap in the Lesser Antilles arc and earthquake hazard in Guadeloupe

Summary. A seismic study of the Lesser Antilles arc has been carried out, first for the period 1950–1978, for which we can use local seismic networks to draw maps of instrumental seismicity, then for the period 1530–1950, for which we have catalogues of felt earthquakes. The striking feature of the spatial distribution of foci is the cluster of epicentres in the northern half of the arc; all large earthquakes ( M > 7.5) are located north of 14° latitude. Seismicity cross-sections through the arc show a variable dipping subduction zone along the arc; the deep seismic zone is steeper in the centre of the arc than on the extremity.
The time-space diagram for historical seismicity, and the evidence of a seismic gap at the east of Guadeloupe lead us to consider the northern half arc as a likely site for a large earthquake in the near future.
The seismic slip rate calculated from all major earthquakes since 1530 is of much greater value than that obtained from recent plate tectonic models, suggesting that the recurrence rate of earthquakes is more than many hundreds of years with a possible aseismic creep.     

THE UNUSUAL STORMS OF FEBRUARY 1992 IN SOUTHERN CALIFORNIA

During February 1992, a series of relatively warm storms passed eastward across southern California, yielding intense precipitation that triggered widespread mass movement, flooding, property damage, and loss of life. These storms were triggered by an intense low pressure system (976 mb) off northern California which deepened as its eastward progress was initially blocked by a high pressure ridge (1040 mb) across western North America. Between February 10 and 13, large areas of Ventura and Los Angeles counties experienced cumulative precipitation of 200–400 mm with intensities reaching 40–50 mm hr^-1. Mass movement, mainly as soil slips that transformed downslope into debris flows, occurred where cumulative precipitation exceeded 300 mm and when sustained intensities exceeded 25 mm hr^-1. Stream response was rapid, particularly in urban areas where impermeable surfaces and storm drains fed concrete stream channels. The canalized upper Los Angeles River and Arroyo Simi exceeded all previous discharges for over 43 and 36 years of record, respectively. Other streams, from the large Santa Clara River to modest Malibu Creek, yielded recurrence intervals for the peak discharge of between 8 and 24 years, but the rapidity of flooding everywhere was remarkable. Whereas main trunk streams, canalized or not, responded predictably, the storm series emphasized the problems of poorly controlled development of potentially unstable hillsides and floodable lowlands and indicated a need to reassess the assumptions upon which such development is permitted. [Key words: climatology, cyclonic storm, geomorphology, mass movement, flooding, California].     

Regional mid-ocean stress and a proposed focal mechanism of 'stress-discordant' earthquakes

Summary. Rock stress measurements in Iceland show maximum horizontal compression perpendicular to the trend of Reykjanes Ridge crest and of its extension, the active volcanic zone of Iceland. Fault-plane solutions of dormant stage earthquakes are consistent with the measured stress orientations, but strike—slip earthquakes associated with volcanic surges and some earthquake swarms in active geothermal areas exhibit apparent reversals of mechanism and are here defined as 'stress-discordant' in the sense that they yield deduced stress orientations 90° from the regional stress field as determined by hydrofracturing and strain relief methods. It is proposed, supported by comparison with the pore-pressure induced Denver earthquakes, that the 'stress-discordant' volcanic earthquakes are triggered by increased pore pressure and probably involve stick-slip motion similar to that reported for some laboratory tests of the pore pressure effect, characterized by gradual onset and sudden stopping of each slip episode. The question is raised as to whether stress-discordant earthquakes are dominated by a stopping phase or terminal shock with consequent reversal of the deduced shear couple. A possible stopping mechanism is suggested: the dilatant stiffening of fault gouge during shear.
It is proposed that direct measurements of stress orientation be made by hydrofracturing tests at other places along the mid-ocean ridge crest and on the margins of the Red Sea and East African rifts. The Icelandic stress data indicate the need for sceptical re-examination of some fundamentals of plate tectonics theory.     

Earthquake swarms as a long-range precursor to large earthquakes in Turkey

Summary. Turkey has been the location of a series of major earthquakes during this century. This study is an attempt to predict these in hindsight using swarms of weak earthquakes as a long-range precursor as proposed by Keilis-Borok. Some modifications of the swarm identification algorithm are made and statistical measures of success to judge the success of the prediction scheme were introduced. The main measures of success are the percentage of large earthquakes predicted and the percentage of swarms that predicted large earthquakes. The method was applied separately to earthquakes in the North Anatolian Fault Zone and in Western Turkey. The North Anatolian Fault was first considered in its entirety and then in segments. Prediction was attempted in each of these regions with a variety of parameters and the measures of success with confidence levels are computed.
The results obtained for prediction in Turkey are promising. The success of predicting large earthquakes ( M ≥ 7) was generally greater than 60 per cent. The difficulties of this method arise from incomplete catalogues of seismicity and the use of many arbitrary parameters.     

Stress drops, wave spectra and recurrence intervals of great earthquakes — implications of the Etorofu earthquake of 1958 November 6

Summary . The great Etorofu earthquake of 1958 November 6 is characterized by a relatively small aftershock area (70 × 150 km²) and an extremely large felt area. The felt area is more extensive than those of any other large earthquakes which have occurred in the southern Kurile to northern Japan arc since the beginning of this century. The mechanism is a pure thrust fault typical of most great earthquakes in island arcs. A body wave magnitude of m _b = 8.2 is obtained at periods around 6 s using more than 40 observations, although an m _b value of only 7.6–7.7 would be expected empirically from the observed surface wave magnitude of M _s= 8.1–8.2. Both an unusually large felt area and a high m _b indicate a dominance of high-frequency components in the seismic waves. A seismic moment of M _o= 4.4 × 10²⁸ dyne cm is determined from long-period surface waves from which a high stress drop of Δσ = 78 bar is obtained using a relatively small aftershock area. Historic data indicate an anomalously long time interval between the 1958 event and any earlier great earthquake from the same source region. The observed high stress drop can be interpreted as a consequence of this long intervening period through which strain built up. The dominance of the high-frequency seismic waves can then be interpreted as a result of this high stress drop. Stress drops, seismic wave spectra and recurrence intervals of great earthquakes are in this way closely related to each other. The 1958 event may represent a high strength extreme of stochastic fluctuation of fracture strength relevant to great earthquakes.     

Spherical harmonic representation of the magnetic field in the presence of a current density

Summary. We present the results of a systematic study of events with M _s > 6 in northern Chile (20–33°S), for the period between 1963 and 1971. Medium to large earthquakes near the coast of this region are of three types: (1) Interplate events at the interface between the downgoing slab and the overriding South American plate. These events can be very large reaching magnitudes greater than 8. (2) Intra-plate earthquakes 20–30 km inside the downgoing slab. They have fault mechanisms indicating extension along the dip of the slab and may have magnitudes up to 7.5. (3) Less frequent, M _s∼ 6 events that occur near the top of the downgoing slab and have thrust mechanisms with an almost horizontal E-W compressional axis. This type of mechanism is very different from that of the events of type 1 which are due to shallow dipping reverse faulting. There is a rotation of about 30° of the compressional axis in the vertical plane between events of types (1) and (3). Three groups of events near 32.5°, 25.5° and 21°s were studied in detail. Depth and mechanisms were redetermined by P -wave modelling and relative locations were obtained by a master event technique. Near 32.5°S, only events of types 1 and 2 were found in the time period of this study. At the two other sites, the three types of events were identified. This shows clearly that there are compressive stresses at the top of the slab and extension at the centre, a situation which is usually found in the areas where a double Benioff-zone has been identified in the seismicity.     

The NW Scotland earthquake swarm of 1974

Summary. A swarm of earthquakes occurred in the Kintail area of NW Scotland in 1974 August with magnitudes up to M_L = 4.6 and reported intensities up to VI MM. One earthquake of the series was recorded by about 60 temporary stations of the LISPB refraction experiment in addition to the permanent LOWNET network. This allowed the determination of an accurate epicentre for that event which was then used as a control event to locate other earthquakes of the Kintail series. The events of that swarm had a strikeslip focal mechanism and seemed to be associated with the Strathconon Fault.     

Moment tensors of microearthquakes from the Eyjafjallajökull volcano in South Iceland

Analyses of relative P - and S -wave amplitudes of 15 selected earthquakes ( M _L <2.3) from a seismic swarm, which occurred in May and June 1994 at the Eyjafjallajökull volcano in South Iceland, reveal similar radiation patterns, a thrust-type double-couple with an additional source component. All focal solutions have nearly vertical T -axes and horizontally oriented P -axes, with E-W-oriented nodal planes. The volume increase corresponding to an isotropic source component is estimated to be in the range of 24 m³. The temporal and spatial seismic pattern, small magnitude range, focal mechanisms and depth range of the Eyjafjallaökull earthquakes indicate vertical intrusion of magma into a confined region at the northern flank of the volcano.     

The influence of water on the stress supported by experimentally faulted Westerly granite

Summary. Fault zones in wet Westerly granite deformed at temperatures of 300° and 400°C require markedly lower shear stresses for sliding than when dry, even when the effective confining pressure is held constant between the wet and dry tests, provided that the strain rate is lower than 10⁻⁷s⁻¹. The rate of strength reduction is enhanced by increasing the pore water pressure. The deformation rate is a power function of the applied stress where the stress exponent is approximately 7 for pore water pressure of 100 MPa and 21 for pore water pressure of 20 MPa.
The experimental results are extrapolated to conditions believed to occur at depths of 10 km along the San Andreas Fault Zone. It is suggested that for slow tectonic deformation at strain rates of 10⁻¹¹ and 10⁻¹⁴s⁻¹ the shear stress for sliding on faults in granite would be approximately 60 and 20 MPa, respectively, at pore water pressures equal to the hydrostatic head. Fluid overpressures of c. 0.8 lithostatic pressure are required to lower the shear stress for sliding into the 10 MPa range at the slower strain rate.     

Sequence of instability processes triggered by heavy rainfall in the northern Italy

Northern Italy is a geomorphologically heterogeneous region: high mountains, wide valleys, gentle hills and a large plain form a very varied landscape and influence the temperate climate of the area. The Alps region has harsh winters and moderately warm summers with abundant rainfall. The Po Plain has harsh winters with long periods of subfreezing temperatures and warm sultry summers, with rainfall more common in winter.Geomorphic instability processes are very common. Almost every year, landslides, mud flows and debris flows in the Alpine areas and flooding in the Po flood plain cause severe damage to structures and infrastructure and often claim human lives. Analyses of major events that have struck northern Italy over the last 35 years have provided numerous useful data for the recognition of various rainfall-triggering processes and their sequence of development in relation to the intensity and duration of rainfall. Findings acquired during and after these events emphasise that the quantity and typology of instability processes triggered by rainfall are related not only to an area's morphological and geological characteristics but also to intense rainfall distribution during meteorological disturbances. Moreover, critical rainfall thresholds can vary from place to place in relation to the climatic and geomorphological conditions of the area. Once the threshold has been exceeded, which is about 10% of the local mean annual rainfall (MAR), the instability processes on the slopes and along the hydrographic networks follow a sequence that can be reconstructed in three different phases.In the first phase, the initial instability processes that can usually be observed are soil slips on steep slopes, mud–debris flows in small basins of less than 20 km² in area, while discharge increases substantially in larger stream basins of up to 500 km². In continuous precipitation, in the second phase, first mud–debris flows can be triggered also in basins larger than 20 km² in area. Tributaries swell the main stream, which is already in a critical condition. The violent flow causes severe problems mainly along valley bottoms of rivers with basins up to 2000 km² in area. First bedrock landslides can occur, reaching a considerable area density, with volumes from a few hundred up to about one to two million cubic meters. In continuous precipitation, in the third phase, basins of more than 2000 km² in area reach their first critical stage. River-bed morphology is extensively modified, with erosional and depositional processes which can locally undermine the stability of structures and infrastructures. Waters overflow levees, flooding villages and towns to various widths and depths and sometimes claiming casualties. Some days after an intense rainfall period, large landslides involving the bedrock can still take place. These processes usually cause the movement of very large rock masses. The total duration of rainfall usually has a greater effect on these landslides than does the number of short periods of very intensive precipitation. This sequence cannot be divided into separate phases when the events occur simultaneously because of the presence of intense rainfall pulses and the generation of very diffuse surface runoff. Such situations usually happen during short-lasting heavy summer rainstorms or in late spring, when snow melt combines with intense rainfall. The three-phase sequence has been identified in three severe events that are analysed in this paper: Valtellina (Lombardy) in 1987, Tanaro Valley (Piedmont) in 1994 and Aosta Valley in 2000; but this sequence has also been observed during other events that occurred in northern Italy: in Piedmont in 1968, 1977, 1978, 1993 and 2000; in Lombardy in 1983 and 1992; in the Aosta Valley in 1993.     

A test of the precursory accelerating moment release model on some recent New Zealand earthquakes

The proposal that the moment release rate increases in a systematic way in a large region around a forthcoming large earthquake is tested using three recent, large New Zealand events. The three events, 1993–1995, magnitudes 6.7–7.0, occurred in varied tectonic settings. For all three events, a circular precursory region can be found such that the moment release rate of the included seismicity is modelled significantly better by the proposed accelerating model than by a linear moment release model, although in one case the result is dubious. The 'best' such regions have radii from 122 to 167 km, roughly in accord with previous observations world-wide, but are offset by 50–60 km from the associated main shock epicentre. A grid-search procedure is used to test whether these three earthquakes could have been forecast using the accelerating moment release model. For two of the earthquakes the result is positive in terms of location, but the main shock times are only loosely constrained.     

广东沿海陆地地质灾害及其防治对策

通过典型地质灾害的调查以及对其控制条件和诱发因素的分析表明 ,广东沿海的强震多出现于NEE活动断裂与NNW、NW向活动断裂交汇部位附近。在工程和经济活动剧烈的丘陵、河流沿岸、交通线和城镇区为水土流失、崩塌和滑坡斜坡灾变的多发区 ,降雨制约斜坡灾变的机率。地裂缝与胀缩土活动相关 ,雨季与旱雨季的交替变化是地裂缝发生的主要原因。岩溶塌陷和地面沉降均与超采地下水有关。三角洲和滨海平原软土区构成了工程建筑非稳定地基。地质灾害防治应以保护自然生态环境为主。丘陵区应进行植树种草 ,涵养水土 ,防止或抑制坡面侵蚀 ;沿海平原区 ,应合理开发利用地下水资源 ,保持地下水动态平衡。同时 ,应合理制定城镇规划 ,对重大工程建设场应开展地质灾害风险评估和地质灾害监测、预报工作。