High Seismicity Rate Research Articles

Active tectonics of the Eastern Sunda and Banda Arcs

The tectonics of the eastern Sunda and Banda arcs of eastern Indonesia are inferred from the centroid depths, seismic moments, and fault plane solutions of 63 large earthquakes that occurred between 1962 and 1984. Source parameters are estimated by inverting long‐period, teleseismic P and SH waves. Earthquakes show that the collision of the Australian continent with the Banda arc shortens the upper plate beneath the forearc and back arc basin over the broad area between the Timor and Seram troughs. Along the eastern end of the Sunda arc and along the southern Banda arc, earthquakes reveal that the arc and forearc respond to collision by shortening in the direction of convergence, by elongating in the direction perpendicular to convergence, and by thrusting over the back arc basin. Shallow thrust and strike‐slip earthquakes beneath the Banda Basin (the back arc basin) demonstrate that deformation within the back arc accommodates some of the northward motion of the Australian continent. The rate of north‐south shortening of the entire upper plate at the longitude of Timor calculated from the seismic moments of the earthquakes over the 22‐year period is roughly 20% of the predicted convergence rate between Australia and southeast Asia. Strike‐slip faulting within the Banda Basin results also in eastward motion of the Banda arc with respect to western Indonesia and Australia. Eastward extrusion of the Banda arc results in thrusting at the Aru trough. Earthquakes due to thrusting of the back arc basin beneath the volcanic arc are too shallow to corroborate that the arc's polarity has reversed although the Wetar back arc thrust zone at present has a high seismic slip rate. Earthquakes and their mechanisms provide evidence that the lithosphere of the Weber Basin, a 7500‐m‐deep forearc basin, rests on the subducting lithosphere, without intervening asthenosphere, so that the Weber Basin subsides in response to sinking of the subducting lithosphere. Closure of the Banda Basin is geometrically incompatible with the Australian plate subducting simultaneously beneath both the Timor and Seram troughs. Instead, the Bird's Head, which subducts beneath Seram, is decoupled from the Australian plate in western New Guinea and probably moves west or southwest with respect to Australia. In the eastern Indonesian example, strike‐slip faulting plays a major role in the evolution of the collision zone and clearly demonstrates the complexity in three dimensions of continental accretion and mountain building by arc‐continent collision.

Microseismicity and focal mechanisms of the intermediate‐depth Bucaramanga Nest, Colombia

The Bucaramanga nest of intermediate‐depth seismicity in northern Colombia was investigated from microearthquakes (mb ≤ 4.3) recorded from June 20 to July 6, 1979, by a local array of 12 digital three‐component seismographs. Of 161 events studied, 142 are from the nest and 19 define a thin Wadati‐Benioff (W‐B) zone, mainly from 110 to 190 km depth. The W‐B zone, part of the Bucaramanga segment, strikes N10°E with a dip increasing with depth from 30°E to near‐vertical beneath the nest. The 89 best located events from within the nest define a volume of about 8 km (NW) by 4 km (NE) by 4 km (depth) centered at 161 km depth, with relative location accuracies of ±0.5 to ±1.0 km. Double‐couple fault plane solutions of 59 events were determined from an objective procedure using P wave polarities and S/P amplitude ratios. There is no dominant trend in focal mechanisms nor any reasonable correlation with regional stress patterns or previously determined mechanisms from teleseismic data. The subduction and assimilation of a linear feature (such as an island arc undergoing a phase transformation) may explain the existence of the Bucaramanga nest. However, the extraordinarily high rate of seismicity suggests that the nest may be moving at a rate several orders of magnitude faster than the rate of relative plate motion alone. The intense seismicity and variation in focal mechanisms suggest that the nest represents some aspect of stress release associated with the ascent of fluids, perhaps brine or magma, that will ultimately result in the initiation of volcanic activity over this recently emplaced W‐B zone.

Analysis of the seismicity detected in 1982-1984 in the Northern Peninsular Ranges of Baja California

The San Miguel fault system strikes in the northwest direction within the Peninsular Ranges of Baja California. The study concentrates primarily on the southern sector of the San Miguel fault (SSM) which is about 20 km long and had a large rupture in 1956 . There is a high rate of seismicity along the SSM. In a time period of 23 days, 109 events were located, with duration magnitudes between 1.4 and 2.8. Focal depths are between 5 and 15 km. Epicentral locations show three clusters of seismicity. Two clusters are on the SSM. Another cluster is closer to a normal fault that strikes parallel 8 km to the southwest of SSM. The seismicity is stationary, with an average of 20 events per day. A “ b ” value of 0.92 ± 0.02 is characteristic for the background seismicity. A larger “ b ” value of 1.4 ± 0.07 is associated with the cluster at the northwest end of SSM. Those inconsistent “ b ” values suggest a heterogeneous state of stress . Most fault plane solutions in the Northern Peninsular Ranges of Baja California indicate strike-slip motions (one normal mechanism is also observed). The greatest principal stress is oriented approximately north-south .

The digital array at Anza, California: Processing and initial interpretation of source parameters

Seismicity along the San Jacinto fault suggests that a 40‐km‐long section near the town of Anza may constitute a gap in the occurrence of ML = 6–7 earthquakes. The potential for a gap‐filling shock and the high rate of seismicity at the southern end of the gap (five events of 4.0 ≤ ML ≤ 5.5 since 1970) provided the impetus for deploying a digital seismic array to collect high‐quality ground motion recordings of all events 2 ≤ ML ≤ 4.5 (on‐scale recording for shocks with magnitudes above ML = 4 can be obtained from an existing array of eight strong motion accelerographs). The Anza site also had the advantage of being in the southern California batholith, which appears to be relatively homogeneous compared to the Franciscan/Gablian contrast of the central San Andreas; we expected that the granitic rocks of the batholith would yield relatively accurate earthquake locations and efficiently propagate high frequencies. The field instrumentation is specifically designed for broadband recording (up to 70 Hz) and high dynamic range (96 dB in the on‐site digitizer alone), since both are necessary for determining the rupture history of earthquakes. Both local VHF and microwave digital telemetry transmit the data from the Anza region to San Diego for computer data logging. In the first 30 months of the array's operation, approximately 292 events have been recorded, located, and processed for source parameters. Most events occur in one of five clusters or in a diffuse zone near the Buck Ridge fault. Two of these clusters are located at right‐stepping en echelon offsets (Coyote Creek‐San Jacinto and San Jacinto‐Hot Springs); two others are directly below and about 8 km west of Anza, respectively. The fifth cluster is just to the northeast of the Hot Springs fault. Although event depths are generally between 11 and 14 km, at the southern end of the Hot Springs fault, depths extend to 18 km; these are some of the deepest strike‐slip earthquakes on the San Andreas system. We calculate source parameters such as the scalar moment and stress drop for the analysis of high‐frequency waves, scaling relations, and earthquake interaction. The largest event recorded thus far had a moment of 4.4 × 1021 dyn cm (M = 3.8) and a stress drop of 55 bars. Both arms and Brune stress drops increase with moment; source radius increases only slowly with moment. The maximum values of both the Brune and arms stress drops increase with depth down to 10 km, remain approximately constant to 14 km, and may decrease below 14 km. The data suggest that stress drops of a group of earthquakes can be related to the strength of the upper crust calculated from frictional and quasi‐plastic flow laws, although individual events may not be related. Stress drops on the San Jacinto fault are high compared to those on the central creeping section of the San Andreas, where stress drops are about 10 bars or less. This observation is consistent with the relative rupture area of events of equal ML (5.5) and may be related to the lithologie differences between the two fault segments and the amount of normal stress compared to shear stress.

Historical seismicity near Chagos: a complex deformation zone in the equatorial Indian Ocean

Over the last twenty years, Chagos Bank has a seismicity rate disproportionate to its supposed intraplate location. Earthquake relocation also shows a high seismicity rate in pre-WWSSN time (1912–1963), with seven events located off of the Central Indian Ridge, including large events in 1912 ( M = 6.8) and 1944 ( M = 7.2). This study uses the moment variance technique, a systematic search for the mechanism which best fits P, PP, SH, Love and Rayleigh amplitudes, to determine the focal mechanisms of two pre-WWSSN earthquakes. A test with a recent event of known mechanism demonstrates that accurate focal parameter determination is possible even when only a few good records are available. Moment variance analysis shows a thrust faulting mechanism for the 1944 event, northeast of Chagos Bank near the Chagos-Laccadive ridge, and a strike-slip focal mechanism for a smaller 1957 event west of Chagos Bank. The 1944 event, one of the largest oceanic “intraplate” earthquakes known (moment 1.4 × 10 27 dyne-cm), indicates that the Chagos seismicity reflects not an isolated occurrence of normal faulting as previously thought, but rather regional tectonic deformation extending northeast of Chagos Bank and including thrust, normal and strike-slip events. This seismicity and previously studied seismicity near the Ninetyeast Ridge and Central Indian Basin suggest a broad zone of deformation stretching across the equatorial Indian Ocean. This zone contains all known magnitude seven oceanic “intraplate” earthquakes not associated with subduction zones or continental margins, suggesting that elsewhere such extensive deformation occurs only along plate boundaries. This study proposes that a slow, diffuse plate boundary extends east from the Central Indian Ridge to the Ninetyeast Ridge and north to the Sumatra Trench. A recent plate motion study confirms this boundary and suggests that it separates the Australian plate from a single Indo-Arabian plate.

Intraplate seismicity and stresses in young oceanic lithosphere

We have conducted a systematic study of oceanic intraplate seismicity in the period 1964–1983 occurring in lithosphere younger than 35 million years. Focal mechanisms determined for 26 events using first motions, P wave modeling and surface wave amplitude radiation patterns, along with 13 previously determined mechanisms, indicate a diversity of faulting styles and stress orientations in young oceanic lithosphere. There is no evidence for a worldwide transition from extension to compression at a specific lithospheric age. Tensional axes of normal faulting events are oriented oblique to the spreading direction and do not show extension in the spreading direction. Congressional axes of thrust events show a weak preferred orientation in the spreading direction, presumably due to ridge push stresses. However, stress axes of thrust events in very young lithosphere (age <9 Ma) do not show a correlation with the spreading direction, suggesting that the ridge push force is small near the ridge. These observations indicate that tensional stress in the spreading direction is concentrated in a very narrow zone at the ridge axis, in agreement with models which assume that the axial region has very little tensile strength. Normal faulting events occur at greater depths and temperatures than thrust events, possibly because rocks under tensional stress have strength maxima at higher temperatures. Alternatively, the depths of the normal faulting events may be a result of the higher thermal gradients at depth in young lithosphere, if thermoelastic stresses are important in generating this seismicity. Differential thermal contraction along fracture zones does not seem to be a significant source of near‐ridge intraplate seismicity. High seismicity rates in young lithosphere, noted previously on a global scale, seem characteristic of all three major ocean basins. Additionally, most of the total oceanic intraplate moment release takes place in young lithosphere. However, most of the events and more than 96% of the total moment release of the near‐ridge intraplate seismicity are contained in five areas. Two zones in the Indian Ocean, one near Chagos Bank and the other in the Amsterdam‐St. Paul region, indicate tension parallel to the trend of the plate boundary. A region of high seismicity near the East Pacific Rise between 3° and 7°N shows both normal and thrust faulting, while another region, on the Cocos plate west of the Panama Fracture Zone, is characterized by strike‐slip faulting. Thrust, strike‐slip, and normal faulting events are found in a small seismically active zone of the South‐Central Pacific. The extremely uneven distribution of seismicity, if not a result of the relatively short time period studied (20 years), indicates local perturbations strongly affect near‐ridge seismicity.

Seismic quiescence and asperities in the Tonga‐Kermadec Arc

Highly significant temporal and spatial seismicity rate changes were found along the TongasKermadec plate boundary. The standard deviate (z) test was used to compare systematically the rates in different periods and different volumes, and to test for uniqueness of anomalies found. The NOAA hypocenter data file showed a pronounced decrease in the number of small‐magnitude earthquakes reported after 1969 at the same time as such a reporting decline exists in the worldwide data. Therefore only earthquakes with mb≥4.9 were studied. A search for quiescence before recent mainshocks yielded two uniquely significant precursory anomalies, one missed mainshock and one false alarm. Quiescence started 63 months before the January 1976, M = 8.0, Kermadec; and 25 months before the June 1977, Tonga, M = 7.2, earthquakes. The December 1975, M = 7.5, shock was not preceded by a seismicity rate change. The seismic gaps in the Tonga‐Kermadec arc with sufficient seismicity for analysis (17°–22° and 31°–34°) show constant rates up to the present. Therefore, on the basis of the quiescence hypothesis, earthquakes are not forecast for these segments. The stress drops, Δσ, of 380 interface events were estimated using the mb/Ms method. The average Δσ was uniform along the island arc with the exception of two segments with average Δσ of less than half the normal values. These segments coincided with the two largest recent interface ruptures; however, in the northern Tonga the Δσ was low before the 1975 rupture, whereas in the Kermadec the 1976 aftershocks showed low Δσ. Asperities could not be identified on the basis of high stress drops. A systematic study of the seismicity rate as a function of space defined large variations along the arc. The areas of highest and lowest seismicity rate correlate with the northern end of the trench and with the Louisville ridge intersection, respectively. However, other large rate contrasts exists between arc segments without obvious tectonic differences. This implies that seismicity doughnut patterns do not reliably define precursors. One segment of outstandingly high seismicity rate (30°–31°S) is identified as a major asperity because it contains a large number of m>6 events and a large number of recent mainshock‐aftershock sequences and because it stopped the great 1976 rupture. We propose that the only other segment of high seismicity rate (near 20.5 °S) also represents a major asperity.

Local seismic activity in the region of the Assam Gap, northeast India

abstract A tripartite array of vertical seismographs was operated for 5.5 months in late 1979 in the Shillong Massif area of northeast India. During this time, 719 local earthquakes were recorded. Of these, 359 were located (2.4 ≦ M D ≦ 5.1), using for some events additional arrival times from the WWSSN/IMD station Shillong and some portable stations operated by the NGRI/RRL. The area of the Shillong Massif and the Mikir Hills (60 km radius within our array) showed a high seismicity rate of about four earthquakes per week with M D ≧ 3 1 4 . This activity rate contrasted strongly with the Assam Gap where there were practically no earthquake sources detected by our array. A seismic asperity seems to exist in the Kopili Valley (between Shillong and Mikir Hills) which may have been the termination of the 1897 rupture, and which could perhaps become the nucleation point for the next great Assam earthquake. Based on these seismicity results and macroseismic data (Seeber and Armbruster, 1981), we estimate the length of the Assam Gap to be about 240 km. Frequency versus magnitude plots revealed b values of about 1.5, and suggested that within the array the locations were complete for M D ≧ 3 1 4 . A comparison of local and teleseismic locations and magnitudes could be made only for the largest event which was located 300 km outside the array. This comparison suggests that M D ≈ m b + 0.5, and that the epicenter location errors outside the array are of the order of 40 km.

Seismic quiescence, stress drops, and asperities in the New Hebrides arc

abstract The rate of occurrence of earthquakes shallower than 100 km during the years 1963 to 1980 was studied as a function of time and space along the New Hebrides island arc. Systematic examination of the seismicity rates for different magnitude bands showed that events with mb < 4.8 were not reported consistently over time. The seismicity rate as defined by mb ≧ 4.8 events was examined quantitatively and systematically in the source volumes of three recent main shocks and within two seismic gaps. A clear case of seismic quiescence could be shown to have existed before one of the large main shocks if a major asperity was excluded from the volume studied. The 1980 Ms = 8 rupture in the northern New Hebrides was preceded by a pattern of 9 to 12 yr of quiescence followed by 5 yr of normal rate. This pattern does not conform to the hypothesis that quiescence lasts up to the mainshock which it precedes. The 1980 rupture also did not fully conform to the gap hypothesis: half of its aftershock area covered part of a great rupture which occurred in 1966. A major asperity seemed to play a critical role in the 1966 and 1980 great ruptures: it stopped the 1966 rupture, and both parts of the 1980 double rupture initiated from it. In addition, this major asperity made itself known by a seismicity rate and stress drops higher than in the surrounding areas. Stress drops of 272 earthquakes were estimated by the MS/mb method. Time dependence of stress drops could not be studied because of changes in the world data set of Ms and mb values. Areas of high stress drops did not correlate in general with areas of high seismicity rate. Instead, outstandingly high average stress drops were observed in two plate boundary segments with average seismicity rate where ocean floor ridges are being subducted. The seismic gaps of the central and northern New Hebrides each contain seismically quiet regions. In the central New Hebrides, the 50 to 100 km of the plate boundary near 18.5°S showed an extremely low seismicity rate during the entire observation period. Low seismicity could be a permanent property of this location. In the northern New Hebrides gap, seismic quiescence started in mid-1972, except in a central volume where high stress drops are observed. This volume is interpreted as an asperity, and the quiescence may be interpreted as part of the preparation process to a future large main shock near 13.5°S.