Minimum 1D Velocity Model from Local Earthquake Data in the Provence Region,South-Eastern France

We computed a one-dimensional (1D) velocity model and station corrections refering to the Provence region (South-eastern France) by inverting P-wave arrival times recorded on an eight-station local seismic network. Using this velocity model and the program HYPOELLIPSE (Lahr, 1989), we relocated a set of 108 local events. The quality comparison between previous earthquake locations and new relocated shows a good improvement.The obtained Minimum 1D velocity model can be used in a better-quality routine for earthquake location and represent a first step towards more detailed seismic studies in the Provence region.     

Foreshock sequence of September 26th, 1997 Umbria-Marche earthquakes

On September 3^rd (22:07 UTC), 1997 a small earthquake with Mw=4.54 started the foreshocks sequence (1500 events with M_L <3.1) of the September 26^th seismic sequence. Two days after, three seismic stations of the University of Camerino were installed around the macroseismic epicenter of the foreshock. We present in this paper the location of foreshocks (with 2.1L<3.3) which occurred between September 3^rd and 26^th. Foreshocks location, with horizontal (ERH) and vertical (ERZ) error less than 1.5 km, define an area 4 km large. Foreshocks have been localized between the epicenters of the two major events of September 26^th, which occurred at 00:33 UTC with Mw=5.6 and at 09:40 UTC with Mw=6.0 (Amato et al., 1998; Ekström et al., 1998). In a vertical cross-section, hypocenters show a low angle (30^°) structure with SW dip-direction. Focal mechanisms for three of the major events show dip-slip fault solutions with strike direction of about N130, in agreement with the CMT solutions of September 3^rd and September 26^th earthquakes (Ekström et al., 1998). Data recorded at two stations Popola (POP) e Capodacqua (CPQ) located on the rupture area of the September 26^th faults, allowed us to calculate a mean Vp/Vs ratio of 1.84±0.03 for the foreshock. This value is lower than the Vp/Vs ratio of 1.89±0.02 calculated for the aftershock sequence occurred in the same area. Besides, the Vp/Vs ratio during the foreshocks sequence is not stable in time but it seems to increase approaching September 26^th. After September 26^th mainshocks, this value tends to stabilize around a higher value of 1.89. Following the dilatancy model, we suggest that the relative low Vp/Vs ratio before the main shocks could indicate the presence of fluid in the focal volume. The presence of fluids could have increased the effective stress on the fault plane and could be responsible for the long foreshock activity before the two main earthquakes of September 26^th. Therefore, we suggest that this foreshock activity could have also contributed to reduce the friction along the September 26^th fault plane, breaking the active structure in two smaller segments. In this hypothesis, foreshock activity could have drastically contributed to mitigate the seismic potential of the Colfiorito's active structure.     

Source characteristics of the 1992 Nicaragua tsunami earthquake inferred from teleseismic body waves

We analyzed the broadband body waves of the 1992 Nicaragua earthquake to determine the nature of rupture. The rupture propagation was represented by the distribution of point sources with moment-rate functions at 9 grid points with uniform spacing of 20 km along the fault strike. The moment-rate functions were then parameterized, and the parameters were determined with the least squares method with some constraints. The centroid times of the individual moment-rate functions indicate slow and smooth rupture propagation at a velocity of 1.5 km/s toward NW and 1.0 km/s toward SE. Including a small initial break which precedes the main rupture by about 10 s, we obtained a total source duration of 110 s. The total seismic moment isM _o =3.4×10²⁰ Nm, which is consistent with the value determined from long-period surface waves,M _o =3.7×10²⁰ Nm. The average rise time of dislocation is determined to be 10 s. The major moment release occurred along a fault length of 160 km. With the assumption of a fault widthW=50 km, we obtained the dislocationD=1.3 m. From andD the dislocation velocity isD=D/0.1 m/s, significantly smaller than the typical value for ordinary earthquakes. The stress drop =1.1 MPa is also less than the typical value for subduction zone earthquakes by a factor of 2–3. On the other hand, the apparent stress defined by 2E _s /M _o , where andE _s are respectively the rigidity and the seismic wave energy, is 0.037 MPa, more than an order of magnitude smaller than . The Nicaragua tsunami earthquake is characterized by the following three properties: 1) slow rupture propagation; 2) smooth rupture; 3) slow dislocation motion.     

The ∼2 ka subplinian eruption of Montaña Blanca,Tenerife

The latest cycle of volcanism on Tenerife has involved the construction of two stratovolcanoes, Teide and Pico Viejo (PV), and numerous flank vent systems on the floor of the Las Cañadas Caldera, which has been partially infilled by magmatic products of the basanite-phonolite series. The only known substantial post-caldera explosive eruption occurred 2 ka bp from satellite vents at Montaña Blanca (MB), to the east of Teide and at PV. The MB eruption began with extrusion of 0.022 km³ of phonolite lava (unit I) from a WNW-ESE fissure system. The eruption then entered an explosive subplinian phase. Over a 7–11 hour period, 0.25 km³ (DRE) of phonolitic pumice (unit II) was deposited from a 15 km high subplinian column, dispersed to the NE by 10 m/s winds. Pyroclastic activity also occurred from vents near PV to the west of Teide. Fire-fountaining towards the end of the explosive phase formed a proximal welded spatter facies. The eruption closed with extrusion of small volume domes and lavas (0.025 km³) at both vent systems. Geochemical, petrological data and Fe-Ti oxide geothermometry indicate the eruption of a chemically and thermally stratified magma system. The most mafic and hottest (875°C) unit I magma can yield the more evolved and cooler (755–825°C) phonolites of units II and III by between 7 and 11% fractional crystallization of an assemblage dominated by alkali feldspar. Analyses of glass inclusions from phenocrysts by ion microprobe show that the pumice was derived from the water-saturated roof zone of a chamber containing 3.0–4.5 wt.% H₂O and abundant halogens (F0.35wt.%). Hotter, more mafic tephritic magma intermingled with the evolved phonolites in banded pumice, indicating the injection of mafic magma into the system during or just before eruption. Reconstruction ot the event indicates a small chamber chemically stratified by in situ (side-wall) crystallization at a depth of 3–4 km below PV. Although phonolite is the dominant product of the youngest activity of the Teide-PV system, there has been no eruption of phonolitic magma for at least 500 years from teide itself and for 2000 years from the PV system. Therefore there could be a large volume of highly evolved, volatile-rich magma accumulating in these magma systems. An eruption of fluorine-rich magma comparable with MB would have major damaging effects on the island.     

Comparison of a complex rupture model with the precursor asperities of the 1975 HawaiiM s-7.2 earthquake

A simplified multiple source model was constructed for the 1975 HawaiiM _s=7.2 earthquake by matching synthetic signals with three component accelerograms at two stations located approximately 45 km from the epicenter. Six major subevents were identified and located approximately. The signals of these are larger by factors of 1.4 to 3.2 than that of theM _L=5.9 foreshock which occurred 70 minutes before the main rupture and also triggered the SAM-1 recorders at the two stations. Dividing the rupture length (40 km) by the duration of strong ground shaking ( 50 sec) an, average rupture velocity of 0.8 km/sec (about 25% of S-velocity) is obtained. Thus it is likely that the rupture stopped between subevents. The approximate epicenters of the 6 major subevents, and of the foreshock, support the hypothesis that they were located in high stress asperities which rupture during the main shock, except for the last events which is interpreted as a stopping phase generated at a barrier. These asperities have been previously defined on the basis of differences in the precursor pattern before the mainshock. Thus, it appears that both the details of the precursors and of the main rupture depended critically on the heterogeneous tress distribution in the source volume. This suggests that main rupture initiation points and locations of high rupture accelerations may be identified before the mainshock occurs, based on precursor anomaly patterns. A satisfactory match of synthetic signals with the observations could be obtained only if the aximuth of the fault plane of subevents was rotated from N60°E to N90°E and back to N30°E. These orientations are approximately parallel to the nearest Kilauea rift segments. Hence the slip directions and greatest principal stresses were oriented perpendicular to the rifts everywhere. From this analysis and other work, it is concluded that this fault surface consisted of three types of segments with different strength: hard asperities (radius 5 km), soft but brittle segments between the asperities (radius 5 km), and a viscous half (10×40 km) which slipped during the mainshock, but where microearthquakes and aftershocks are not common.     

Q measurements for PhaseX overtones

Linear stacking procedures are used to retrieve the attenuation of 91 modes belonging to the 3rd, 4th and 5th Rayleigh overtones branches in the 80–160 s period range, and contributing to the so-called PhaseX wave group. Our data show in general slightly less attenuation than expected from available models. Data space inversion shows that, when combined with previously measured fundamental modeQ's, this new dataset improves resolution significantly in the 1000–2000 km depth range. Based on this remark, we carry out a number of parameter space inversions. Our results suggest a narrow (80–200 km) zone of high attenuation (Q =75–90), low attenuation in the intermediate mantle (670–1500 km); (Q 350), and lower values in the deeper mantle (Q 200).     

Earthquake strong motion simulated by stress drop models

Shallow strike slip earthquakes on vertical faults are modelled as two-dimensional antiplane strain ruptures in a uniformly prestressed homogeneous halfspace. Behind the rupture front, which is specified, the stress drops to a lower value. The elastodynamic boundary value problem is solved with a finite difference method. Several cases are studied, which include symmetric and one-directional rupture propagation, surface faulting, multiple events, variable rupture velocity, sticking and rebreaking of the fault plane. The time function of displacement, velocity and acceleration are interpreted as signals generated by events in the focus, namely starting, stopping and breaking through the surface of the rupture. The model explains peak velocity and peak acceleration in the near field of M5.5–6 earthquakes; which are typically about 0.2 m/s and 5 m/s² at 10 km epicentral distance, if the rupture velocity is close to the shear wave velocity. Sticking of the fault does not alter the accelerograms significantly, but it increases the seismic moment in simple events and decreases it in multiple events.Contribution No. 226, Geophysical Institute, University of Karlsruhe.     

A statistical model for estimating the accuracy of event location applied to a network of Scandinavian stations

A statistical model for estimating the errors of epicenter location and origin time is proposed and applied to the old Fennoscandian seismic network. An average crustal model (Sellevoll andPomeroy, 1968) and P and S wave residuals as a function of azimuth have been used. The calculations are carried out for different maximum detection ranges. The analysis shows relatively, small (1 km) standard errors of epicenter location of strong earthquakes for central Fennoscandia. The largest errors are found in the southern and eastern parts of Fennoscandia.On leave from Institute of Geophysics, Polish Academy of Sciences, Warsaw, Poland.     

云南腾冲地区三维地震定位

本文利用中国数字测震台网和流动台站的地震资料,基于参数优化的AICD自动拾取算法和质量评估方案得到了高质量的震相到时,并在此基础上使用一维、三维定位方法对腾冲地区的799次地震事件进行了重新定位。定位结果显示:水平方向上,一维、三维重定位结果相差较小;深度方向上,三维定位的震源成丛分布比一维定位结果更加密集,地震主要位于地壳内低速层之上。分别利用一维、三维定位方法对典型地震、人工震源进行定位,结果表明,三维定位的精度明显优于一维定位,其在水平、深度方向上的平均绝对定位误差分别为0.7 km和1.3 km。      

Contrasting compositions of Mariana Trough fallout tephra and Mariana Island arc volcanics: a fractional crystallization link

Discrete Quaternary (<400 ka) tephra fallout layers (mostly <1 cm thick) within the siliceous oozes of the central Mariana Trough at 18°N are characterized by medium-K to high-K subalkalic volcanic glasses (K₂O=0.8–3.2 wt.%) with high large-ion lithophile elements (LILE)/high-field-strength elements (HFSE) ratios and Nb depletion (Ba/La35; Ba/Zr3.5; La/Nb4) typical for convergent margin volcanic rocks. Compositional zoning within layers ranges from basaltic to dacitic (SiO₂=48–71 wt.%; MgO=0.7–6.5 wt.%); all layers contain basaltic andesites. The tephra layers are interpreted as single explosive eruptive events tapping chemically zoned reservoirs, the sources being the Mariana arc volcanoes (MAV) due to their proximity (100–400 km) and similar element ratios (MAV: Ba/La=36±7; Ba/Zr=3.5±0.9). The glasses investigated, however, contrast with the contemporaneous basaltic to dacitic lavas of the MAV by being more enriched in TiO₂ (1.2 wt.%; MAV0.8 wt.%), FeO^* (10 wt.%, MAV8–9 wt.%), K₂O (1.1 wt.%; MAV0.8 wt.%) and P₂O₅ (0.4 wt.%; MAV0.2 wt.%). (Semi-)Incompatible trace elements (including Rare Earth Elements (REE)) of the basaltic-andesitic and dacitic glasses match those of the dacitic MAV lavas, which became enriched by fractional crystallization. Moreover, the glasses follow a tholeiitic trend of fractionation in contrast to MAV transitional trends and have a characteristic P₂O₅ trend that reaches a maximum of 0.6 wt.% P₂O₅ at 57 wt.% SiO₂, whereas MAV lavas increase linearly in P₂O₅ from 0.1 to 0.3 wt.% with increasing silica. Both explosive and effusive series are interpreted to have evolved in common magma reservoirs by convective fractionation. Similar parental magmas are suggested to have separated into coexisting Si-andesitic to dacitic and basaltic melts by in situ crystallization. The differentiated melt is interstitial in an apatite-saturated crystalline mush of plag+px±ox±ol at the cooler chamber margins in contrast to the less differentiated basaltic to basaltic-andesitic magmas, which are not yet saturated in apatite and occupy the chamber interior. Reinjection of interstitial melt into the chamber interior and mixing with larger melt fractions of the interior liquid (mixing ratios about 1: 8–9) can explain the paradoxical behavior of apatite-controlled P and MREE variation in the basaltic andesite glasses and their MAV dacite-like fractionation patterns. The process may also account for the exclusively tholeiitic trend of fractionation of the glass shard series, but in situ crystallization alone cannot cause their absolute enrichment in (semi-)incompatible elements. The newly mixed melt is suggested to form the basaltic end member of the glass shard series. However, it must have become physically separated from the main MAV magma body (possibly by density-driven convective fractionation) in order to allow for further evolution of the contrasting geochemical paths as well as differentiation.     

Analysis and automatic processing in near-field of eight 1992–1994 tsunamigenic earthquakes: Improvements towards real-time tsunami warning

We study eight tsunamigenic earthquakes of 1992–1994 with data from single near-field 3-component long-period stations. The analysis is made from the standpoint of tsunami warning by an automatic process which estimates the epicentral location and the seismic moment through the variable-period mantle magnitudeM _m . Simulations of early warning based on the real-time computation of the seismic moment are also tested with this system, which would give a justified warning in each region of tsunami potentiality. By exploiting the dependence of moment rate release with frequency, the system has the capability of recognizing both tsunami earthquakes such as the 1992 Nicaragua and 1994 Java events, as well as instances of the opposite case of low-frequency deficiency, interpreted as indicating a deeper than normal source (1993 Guam event). We report both the results of delayed-time processing of the near-field stations, and the actual real-time warnings at PPT, which confirm the former.     

Unified approach to amplitude attenuation and coda excitation in the randomly inhomogeneous lithosphere

A unified model is proposed for explaining the frequency dependent amplitude attenuation and the coda wave excitation on the basis of the single scattering process in the randomly inhomogeneous lithosphere. Adopting Birch's law and a direct proportion between density and wave velocity, we statistically describe the inhomogeneous medium by one random function characterized by the von Karman autocorrelation function. We calculate the amplitude attenuation from the solid angle integral of scattered wave energy on the basis of the Born approxiimation after subtracting the travel-time fluctuation effect caused by slowly varying velocity inhomogeneities. This subtraction is equivalent to neglect energy loss by scattering within a cone around the forward direction. The random inhomogeneity of the von Karman autocorrelation function of order 0.35 with the mean square fractional fluctuation of 7.2×10^–3 1.3×10^–2 and the correlation distance of 2.15.1 km well explains observed backward scattering coefficientg and the ratioQ _P ^–1 /Q _S ^–1 , and observed and partially conjecturedQ _S ^–1 for frequencies between 0.5 Hz and 30 Hz.     

Estimates of coda Q attenuation in the João Câmara area (Northeastern Brazil)

S coda wave of seventy-four local earthquakes recorded in a network of ten seismic stations were used to calculate coda Q attenuation (Q_c) in the João Câmara area (northeastern Brazil). The estimates show Q_c as a strong function of frequency in the range from 6.0 to 20.0 Hz. We found out that Q_c in João Câmara has a functional form given by Q_c= Q₀ f, where Q₀= 151 ± 99 and = 0.98 ± 0.05. If the standard deviations are taken into account,we conclude that there are no relevant changes in both Q₀ and values from one station to another. The estimated Q₀ values at the different stations suggest that the Samambaia fault is a boundary between two different seismic attenuation zones. In one side of the fault (left), where stations were installed in Pre-Cambrian terrain and thick sedimentary layer, the seismic attenuation is stronger than in the other side (stations installed in thin sedimentary layer and limestone outcrop).The anomalous Q₀ values in the left side of the Samambaia fault can be explained due to the presence of a shallow conductive layer in the upper crust( 10 km), such as proposed by Padilha et al. (1992). According to our results, if there is a conductive layer in the area, it probably spreads over João Câmara city and surrounding regions.However, more detailed investigation either with seismic methods (seismic attenuation,3D tomography with P and/or S wave velocities) or with other geophysical methods is needed to interpret the observed differences in Q₀ values between the two sides of the Samambaia fault.     

A site effect study in the Verchiano valley during the 1997 Umbria-Marche (Central Italy) earthquakes

Strong site effects were observed during the two M _W 5.7 and M _W 6.0 main shocks of the Colfiorito seismic crisis which occured on September 26, 1997 in Umbria-Marche (Central Italy).The most obvious indications of these effects are the dramatic differences in damage shown by buildings of similar construction in neighboring villages.Such observations were specifically made in the Verchiano valley in the fault area, 15 km south of Colfiorito where the Verchiano village and the Colle and Camino hamlets were heavily damaged (MCS intensity IX-X) since the first main shock of 1997/09/26,while in contrast, the Curasci village located 2 km eastwards remains almost intact.In order to study the anomalous ground motion amplifications in this area, an array of 11, 3-components seismo- and accelero-meters was set up during the 1997/10/20-24 period, extending from the western side of the valley, up to the top of Mount San Salvatore, going accross the Colle and Curasci hamlets.During the experiment, 67 aftershocks enlightened the valley from the Colfiorito (10 km north) and the Sellano (6 km south) active swarms.Seismic refraction experiments were conducted at the same time in the 500 m wide, 1500 m long Verchiano valley in order to determine the thickness and main characteristics of the alluvial infilling.The main results are: (i) compared to the valley side ground motion, and for all the events, recordings in the central part of the valley (piana di Verchiano) show relative amplification of 10 with a clear lengthening of the seismogram duration by a factor of 2 – (ii) broad band relative amplification of 6–8 is also clearly identified at the top of the Mount San Salvatore overhanging the valley – (iii) any of the site effect measurements done explains by itself the strongly contrasted damage observed at Colle and Curasci: i.e. the modification of the near-field radiation pattern by interaction with the free heterogeneous surface may have induced local shadow zones that saved Curasci.     

Anomalously deep earthquakes in northwestern Italy

It is usually assumed that earthquakes in intraplate regions occur in the upper crust, and northwestern Italy is generally assigned to this kind of normal seismicity. In this work, the depth distribution of the events localized in this area by the Istituto Geofisico Geodetico (IGG) seismic network in the period 1991–1997 is analyzed in detail. In particular, the location capability of the network is discussed, adopting as reference quarry blasts (for the epicentral position) and the locations obtained from a dense temporary network (for the depth estimate). Within the so-obtained error limits, the depth distribution of events show a characteristic pattern: while for most of the area covered by the network the well-located seismicity lies within the first 20 km of depth, in a band following the inner arc of the Western Alps, numerous events have anomalously large focal depths, reaching a maximum of 114 km. These depth determinations cannot be attributed to instabilities of the location procedure: different choices of the propagation models used for the hypocentral determination led to very similar depth values, always significantly larger than the standard values for the surrounding areas. A strong correlation has been found between the 3-dimensional distribution of these foci and the P-wave propagation anomalies obtained from tomographic studies, suggesting a direct link between elastic and rheological properties of lower crust and upper mantle in this area.     

Scaling vesicle distributions and volcanic eruptions

Models of coalescence-decompressive expansion of the later stages of bubble growth predict that for diverse types of volcanic products the vesicle number densities (n(V)) are of the scaling form where V is the volume of the vesicles and B₃ the 3-dimensional scaling (power law) exponent. We analyze cross sections of 9 pumice samples showing that over the range of bubble sizes from 10 m to 3 cm, they are well fit with B₃0.85. We show that to within experimental error, this exponent is the same as that reported in the literature for basaltic lavas, and other volcanic products. The importance of the scaling of vesicle distributions is highlighted by the observation that they are particularly effective at packing bubbles allowing very high vesicularities to be reached before the critical percolation threshold, a process which—for highly stressed magmas—would trigger explosion. In this way the scaling of the bubble distributions allows them to be key actors in determining the rheological properties and in eruption dynamics.Editorial responsibility: D. Dingwell     

An experimental study of triggered stick-slip

Stick-slips have been studied in the laboratory on granite, labrodorite and sandstone samples of two different sizes. Different roughness was achieved on the sawcut surfaces by finishing them with different grinding compounds ranging from grit 40 to grit 1000. Stick-slips occurred as a result of 1) slowly increasing the shear and normal stresses, 2) superimposing a sinusoidal stress modulation (0.1 and 10 Hz) on the slowly increasing stresses, 3) triggering by a stress impulse when the shear stress was well below the levels where stick-slips occurred without the impulse, and 4) foreshocks.Stick-slips triggered by impulses or foreshocks occurred long after the beginning of the triggering events, i.e., long in comparison with elastic wave travel times through the sample. All triggered events were very rich in high frequencies (corner frequency of 100 kHz). The untriggered stick-slips did not contain much energy at the high frequencies (corner frequency of 10 kHz). The dynamic friction coefficients for the triggered stick-slips were smaller than for the untriggered events.The long delay between the onset of the trigger and the stick-slip, and the high frequencies may be a consequence of corrosion of asperities. The ultimate triggering and the rate of corrosion are likely related to the interplay of the normal and shear stresses as they load and unload the fault surface. The consistent shape of the high frequency spectra is probably due to sample resonances which are excited rather than being characteristic of the details of the stick-slips. If these laboratory observations are directly applicable to earthquake seismology, the spectra of earthquakes which were triggered by other earthquakes should be anomalously rich in high frequencies.     

Microearthquake activity associated with underground coal-mining in Buchanan County,Virginia, U.S.A.

Microearthquake activity (impulsive, transient seismic events, with durations up to several seconds at a distance of 500 m, that exhibit a coda with a shift toward lower frequencies with increasing time) was monitored for a three-month period by a single seismograph sited directly above an undergound longwall mine in the coal-mining region of Buchanan County, Virginia, U.S.A. The purpose of this investigation was to determine if precursory increases in microseismicity prior to cavings (subsidence) of overburden in the mine were present and, if so, could they be detected by surface seismographic observations. The first two recording weeks were prior to the beginning of coal removal operations at the monitored mine. A comparision of the before and after levels of microearthquake occurrence indicated a sevenfold increase to about seven seismic events/hour that was attendant with the development of the time over the level of the background, non-coal-mining period seismicity.A total of over 15,000 microearthquakes were recorded during the monitoring period, most of which occurred during the actual coal-mining operations. The workday rate exceeded 30 seismic events/hour in contrast with the non-workday rate of about seven such events/hour. Rock and coal fracturing ahead of the mine plow are believed to be the primary cause of the majority of these very small seismic events. Cavings and rockbursts (violent eruptions that propel rock debris into the mine) also contributed to the total seismic activity. It appears that cavings, some of which were large enough to be felt on ground surface, are the primary source of the non-plowing related seismicity as larger free surface areas are opened underground. Any seismic activity premonitory to cavings, however, was effectively masked by the high workday rate. Thus, the use of surface seismic monitoring, in an attempt to document any increases of localized seismicity precursory to cavings, failed in this instance.The exact location of the mine and the survey dates are not given in this paper at the request of the mine operator.     

Comparison of Integrated Geodetic Data Models and Satellite Radar Interferograms to Infer Magma Storage During the 1991–1993 Mt. Etna Eruption

We compare the results obtained from the modelling of EDM, GPS, levelling and tilt data measured in the first part of the 1991–1993 eruption at Etna to the InSAR data acquired during the second part. The geodetic changes are very marked in the first half of the eruption and constrain a deflation source located at a few kilometers of depth ( 3 km b.s.l.), in agreement with other independent geophysical evidence. SAR data, available during the second part of the eruption, were analysed for different time intervals in the second part of the eruption. The interpretation of SAR interferograms reveals a large-scale but less marked deflation of the volcano that could be caused by a deeper source. This second source is in accord with a second deeper anomaly revealed by recent seismic investigations. The combination of geodetic data modelling and SAR images suggests a complex plumbing system composed at least of two possible storage regions located at different depths.     

Mining-induced microseismic event location errors: Accuracy and precision of two location systems

The accuracy and precision of microseismic event locations were measured, analyzed, and compared for two types of location systems: anolog and digital. In the first system, relative times of first arrival were estimated from analog signals using automated hardware circuitry; station positions were estimated from mine map coordinates; and event locations were determined using the BLD (Blake, Leighton, and Duvall) direct solution method. In the second system, arrival times were manually measured during interactive displays of digital waveforms; station coordinates were surveyed; and the SW-GBM (Salamon and Wiebols; Godson, Bridges, and McKavanagh) direct basis function was used to solve for locations. Both systems assume constant isotropic seismic velocity of slightly different signals data sets, calibration blast signals with known source site and origin time, and microseismic event signals, were recorded by each location system employing the same array of high-frequency (5 kHz) accelerometers with 150 m maximum dimension. The calibration blast tests indicated a location precision of ±2 m and accuracy of ±10 m for the analog system. Location precision and accuracy for the digital system measured ±1 m and ±8 m, respectively. Numerical experiments were used to assess the contributions of errors in velocity, arrival times, and station positions on the location accuracy and precision for each system. Measured and estimated errors appropriate to each system for microseismic events were simulated in computing source locations for comparison with exact synthetic event locations. Discrepancy vectors between exact locations and locations calculated with known data errors averaged 7.7 and 1.4 m for the analog and digital systems, respectively. These averages are probably more representative of the location precision of microseismic events, since the calibration blast tests produce impulsive seismic arrivals resulting in smaller arrival-time pick errors in the analog system. For both systems, location accuracy is limited by inadequate modeling of the velocity structure. Consequently, when isotropic velocity models are used in the travel-time inversions, the increased effort expended with the digital location system does not, for the particular systems studied, result in increased accuracy.