Lg Coda Research Articles

A Study of Lg Coda Attenuation in North China— the Measurement of Lg Coda Q

AbstractBy using Lg coda time series with high signal‐to‐noise ratio from North China telemetry and digital seismic network during the period from 1989 to 1998 (ML ≥ 3.0), the spatial distribution of Lg coda Q(Q = Q0fn)is obtained with the stacked spectral ratio method through pretreatments of filtering and eliminating background noise. The results show that there is a significant lateral heterogeneity on Lg coda Q distribution in North China, which is approximately consistent with the velocity distribution in the crust of the region, and is related with the surface geological features. High Q0 values appear on propagating paths located in uplift areas, and low Q0 values in depression areas. Middle Q0 values exist on propagating paths between uplift and depression areas. These indicate that the crustal structure beneath the North China region has no abrupt change which can destroy seismic wave‐guide. The Lg coda Q0 values reflects basically intrinsic visoelastic attenuation of the crustal media.

Q of the Indian Shield

We analyze spectral attenuation of Lg waves to determine Q of the Indian shield. The dataset consists of four earthquakes recorded in the distance range of 240-2400 km. The new estimate, Q(f ) 800f 0.42 (0.1 Hz f 20 Hz), is based on a larger dataset than a previous estimate, Q(f ) 508f 0.48 (1 Hz f 20 Hz). The new Q(f ) of the Indian shield is comparable to the Q in regions of eastern North America where its value is relatively high, for example, the Adirondack Mountains. It is also in agreement with previously reported Lg coda Q at f 1 Hz for the Indian subcontinent. The revised Q(f ) has important implications for the estimation of ground motions in the Indian shield region using the stochastic method since the recorded data, presently, come mostly from large distances.

Attenuative Dispersion of P Waves and Crustal Q in Turkey and Germany

We have measured group delays of the spectral components of high-frequency P-waves along two portions of the North Anatolian Fault Zone (NAFZ) in Turkey and in a region of southern Germany. Assuming that the observed dispersion is associated with attenuation in the crust and that it can be described by a continuous relaxation model, we obtained Q and the high-frequency relaxation times for those waves for each of the three regions. Individual P-wave Q values exhibit large scatter, but mean values in the NAFZ increase from about 25 to 60 over the distance range 5–90 km. Mean Q values are somewhat higher in the eastern portion of the NAFZ than in the western portion for measurements made at distances between 10 and 30 km. P-wave Q values in Germany range between about 50 and 300 over the hypocentral distance range 20–130 km. In that region we separated the effects of Q for basement rock (2–10 km depth) from that of the overlying sediment (0–2 km depth) using a least-squares method. Q varies between 100 and 500 in the upper 8–10 km of basement, with mean values for most of the distance range being about 250. Q in the overlying sediments ranges between 6 and 10. Because of large scatter in the Q determinations we investigated possible effects that variations of the source-time function of the earthquakes and truncation of the waveform may have on Q determinations. All of our studies indicate that measurement errors are relatively large and suggest that useful application of the method requires many observations, and that the method will be most useful in regions where the number of oscillations following the initial P pulse is minimized. Even though there is large scatter in our Q determinations, the mean values that we obtained in Turkey are consistent with those found in earlier studies. Our conclusions that Q is significantly higher in the basement rock of Germany than in the basement rock of Turkey and that Q is lower in western Turkey than in eastern Turkey are also consistent with results of Q studies using Lg coda.

Distribution Characteristics of Lg Coda Q in Yunnan and Its Adjacent Regions

AbstractWe have collected broad band seismograms of the earthquakes with magnitudes greater than 4.0 recorded by the Kunming station for the period 1990~2000 and 23 Kunming digital network stations located in Yunnan province for the period 1999~2000. The total 280 Lg coda traces are studied using the stack spectral ratio (SSR) method. Supposing Q of Lg coda correlates with frequency f by Q = Q0fη, the Q value at 1Hz (Q0) and frequency‐related factor η for each record are calculated. Based on the calculated results, the distribution maps of Q0 and η for 0.5°×0.5° cells are obtained using the back‐projection technique. Our results indicate that Q0 in Yunnan and its adjacent regions exhibits low values, varying between 150 and 300. The Q0 in eastern Yunnan is higher than that in western Yunnan. The η varies between 0.3 and 0.8. For most of the regions under study η is within the range from 0.5 to 0.6.

Separation of intrinsic absorption and scattering attenuation fromLgcoda decay in central France using acoustic radiative transfer theory

Radiative transfer theory is applied to the measurement of average values of the transport mean free path and absorption length (l a , related to the intrinsic quality factor) in the lithosphere beneath France by comparingsimulated and observed coda envelopes of regional records. The lithosphere is modelled as a flat layer representing the crust overlying a half-space representing the mantle, with different velocities and scattering properties. We infer the average values of I a (assumed to be constant over the whole medium) and transport mean free paths in the crust (l * c ) and mantle (l * m ) by minimizing a misfit function between observed and modelled Lg coda envelopes. This comparison is conducted in two different time windows, the whole Lg coda (group velocity smaller than 2.6 km s - 1 ) and the early coda (v g < 2.6 km s - 1 and lapse time smaller than 300 s), when coda envelopes have decay rates that depend on distance. In the early coda, a single set of parameters l a and l* c minimizes the misfit function. Using the whole Lg coda time window we find that many models fit the data equally well. We explain this difference by the common decay rate of envelopes independent of distance over most of the coda window. The measurement of the attenuation of the coherent Lg wave complements these results to constrain l a and l* c . We find that a model with l* c ∼ 250 km and l a ∼ 150 km explains the attenuation of both the direct Lg wave train and the entire Lg coda with both time and distance at 3 Hz. However, the trade-off between the two parameters is too strong to conclude reliably on the dominance of intrinsic on scattering attenuation.

Distribution of Lg coda Q in the Chinese continent and its adjacent regions

Distribution of Lg coda Q in the Chinese continent and its adjacent regions

Distribution of Lg coda Q in the Chinese continent and its adjacent regions

785 traces of vertical components from shallow earthquakes recorded by 10 CDSN (Chinese Digital Seismographic Network) stations and 5 GSN (Global Seismographic Network) stations were collected to study the attenuation characteristics ofL g coda in the Chinese continent and its adjacent regions. The records were processed first using the stack spectral ratio method to obtain the average values ofQ 0 (Q at 1Hz) and η, the frequency dependence, ofL g coda in the ellipses corresponding to the paths. The back-projection technique was then employed to obtain the tomographic maps ofQ 0 and η values, and the distribution of their errors. Results indicate that in the studied areaQ 0 varies between 200 and 500. The lowest value ofQ 0 exists in the Yun-nan-Tibetan region, while the highest value ofQ 0 occurs in the southern edge of Siberian platform. η varies between 0.3 and 0.8. For most part of the studied area η varies inversely withQ 0.

Source Characteristics and Propagation Effects of the Puebla, Mexico, Earthquake of 15 June 1999

The analysis of arrival times indicates that the focus of the 15 June 1999 (20:42 UTC; 6.3 m b; 6.5 M s; 7.0 M w) Puebla, Mexico, earthquake is located at 18.15° N and 97.52° W and is 68 km deep in the contact region between the continental American and the Cocos subducting plate. First-motion data and regional and teleseismic waveform modeling indicate that the main rupture occurred as normal faulting along a fault striking N64° W and dipping approximately 42° toward the north-northeast. The aftershock activity, though scarce, is concentrated mainly toward the west of the mainshock. Three successive subevents are necessary to fit observed and synthetic waveforms. The location in space and time of subevents indicates that the rupture started as a small release at 76 km deep and then propagated toward the northwest releasing most of the seismic energy at 73 km deep. Teleseismic spectral analysis indicates that the total seismic moment release was 2.0 × 1019 N-m, and the mainshock fault length is of the order of 18 km with a stress drop of 15 bars. The source-time function consisted of three sources with a total duration of 10 sec. Recorded acceleration amplitudes of the mainshock show a clear enhancement that can be up to eight times higher toward the northwest of the epicenter in the frequency range from 0.06 Hz to 6.0 Hz, compared with the opposite direction. The distribution of reported seismic intensity also shows this feature. The directivity effect cannot be fully explained by the orientation of the fault, neither by the rupture propagation direction; therefore, we investigate the possibility that the directivity could be due to a regional propagation effect by analyzing the Lg coda decay of the mainshock. The average frequency-dependent Q Lg quality factor for the vertical component in the frequency range from 0.2 Hz to 2 Hz was Q Lg( f ) = 200 ± 12 f 0.60±0.17 and Q Lg( f ) = 202 ± 7 f 0.53±0.09 for paths toward the northwest and southeast of the epicenter, respectively. The resulting average factor Q o and η do not show significant azimuthal fluctuations. Therefore, variations of crustal attenuation properties cannot explain the notable extent of felt area and observed damage northwest of the epicenter. Probably the presence to the north of the Transmexican Volcanic Belt has played an important role in the enhancement of low-frequency seismic signals.

Seismic Wave Propagation and Scattering in Heterogeneous Crustal Waveguides Using Screen Propagators: I SH Waves

The great advantages of one-way propagation methods, such as the gen- eralized screen propagators (GSP) method, are the fast speed of computation, often several orders of magnitude faster than the full-wave finite difference and finite element methods, and the huge savings in internal memory. In this article, a half- space GSP is formulated for the SH half-space problem. Two versions of the half- space GSP are derived: the wide-angle pseudo-screen and the phase-screen. The Moho discontinuity is treated as parameter perturbations from the crustal back- ground. The validity and limitations of this treatment are discussed. It is shown that half-space screen propagators can accurately propagate guided crustal waves that are composed of small-angle waves with respect to the horizontal direction. Comparisons of numerical results with a wavenumber integration method for flat crustal models and a finite difference algorithm for heterogeneous models show excellent agree- ments. For a model with propagation distance of 250 km, dominant frequency at 0.5 Hz, the GSP method is about 300 times faster than a finite difference algorithm with a similar accuracy. These comparisons demonstrate the accuracy and efficiency of the method. We apply our method to simulate regional wave propagation in different types of complex crustal waveguides including those with small-scale random het- erogeneities. The influence of these heterogeneities on Lg amplitude attenuation and Lg coda formation is significant.

Lg Coda Q Variations across South America and their Relation to Crustal Evolution

Nine broadband seismograph stations in South America have provided 389 recordings of Lg coda with paths that cover most of the continent. Lg coda Q (Q0v) and frequency dependence values at 1 Hz, obtained from these records, were inverted using back-projection tomography to obtain regionalized maps of Q0 and . The entire western margin of the continent (the active Andean mountain belt) is typified by low Q0 (250–450), whereas broad regions of high Q0 (700–1100) span the central Brazilian shield and contiguous regions to the north and south. Intermediate Q0 (450–700) characterizes the northern Patagonia platform and most of the Atlantic shield. Reduced Q0 in the Atlantic shield may be related to tectonic or igneous activity that occurred during the breakup of Gondwanaland during the Jurassic period. This Q distribution is generally consistent with earlier studies where Q0 was found to be directly proportional to the time that has elapsed since the most recent episode of major tectonic or orogenic activity in any region. Reduced Q0 in the Patagonian platform may, however, be due to young sediments there.

Lg Coda Q in Australia and its Relation to Crustal Structure and Evolution

We have determined Lg Coda Q Lg c ) from ground motion recorded at seven broadband stations in Australia, using a stacked spectral ratio method. In spite of the relatively small number of events and less than optimum station coverage, we were able to use those data to obtain a tomographic map Q Lg c and its frequency dependence, at 1 Hz for almost the entire island continent.

Lg Coda Q and its Relation to the Structure and Evolution of Continents: A Global Perspective

Tomographic maps of Lg coda Q (Q Lg c ) variation are now available for nearly the entire African, Eurasian, South American, and Australian continents, as well as for the United States. Q Lg c at 1 Hz (Q 0) varies from less than 200 to more than 1000 and Q Lg c frequency dependence (η) varies between 0.0 and nearly 1.0. Q 0 appears to increase in proportion to the length of time that has elapsed since the most recent major episode of tectonic or orogenic activity in any region. A plot of Q 0 versus time since that activity indicates that a single Q 0-time relation approximates most mean Q 0 values. Those that deviate most from the trend lay in Australia, the Arabian Peninsula, and the East African rift. The increase in Q 0 with time may be due to a continual increase in crustal shear wave Q (Q μ) caused by the loss of crustal fluids and reduction of crustal permeability following tectonic or orogenic activity. Extrapolated values of Q Lg c at 5 Hz (using Q 0 and η values measured at 1 Hz and assuming that η is constant in all regions between 1 and 5 Hz) show a similar percentage-wise increase with times that has elapsed since the most recent activity. Other factors that can reduce Q 0 in continental regions include thick accumulations of sediment (especially sandstone and shale of Mesozoic age and younger), severe velocity gradients at the crust-mantle transition and, possibly, lateral variations in the depth, thickness, and severity of those gradients. Severe and large increases of Q μ in the mid-crust of some regions can cause relatively large values of η, even if the frequency dependence of Q μ is small.

On the nature of regional seismic phases-III. The influence of crustal heterogeneity on the wavefield for subduction earthquakes: the 1985 Michoacan and 1995 Copala, Guerrero, Mexico earthquakes

The most prominent feature of the regional seismic wavefield from about 150 to over 1000 km is usually the Lg phase. This arrival represents trapped S-wave propagation within the crust as a superposition of multiple reflections, and its amplitude is quite sensitive to the lateral variation in the crust along a propagation path. In an environment where the events occur in a subduction zone, such as the western coast of Mexico, quite complex influences on the character of the regional wavefield arise from the presence of the subduction zone. The great 1985 Michoacan earthquake (MW=8.1), which occurred in the Mexican subduction zone, was one of the most destructive earthquakes in modern history and its notable character was that at Mexico City, located over 350 km from the epicentre, there was strong ground shaking almost comparable to that in the epicentral region that lasted for several minutes. Considerable effort has been expended to explain the origin of the unusual observed waves that caused the severe damage in the capital city during the destructive earthquake. The nature of the propagation process in this region can be understood in part by using the detailed strong-motion records from the 1995 Copala, Guerrero (MW=7.4) earthquake near the coast to the south of Mexico City, which also had an enhanced amplitude in the Valley of Mexico. Numerical modelling of both P and S seismic waves in 2-D and 3-D heterogeneous crustal models for western Mexico using the pseudospectral method provides direct insight into the nature of the propagation processes through the use of sequences of snapshots of the wavefield and synthetic seismograms at the surface. A comparison of different models allows the influences of different aspects of the structure to be isolated. 2-D and 3-D modelling of the 1985 Michoacan and 1995 Copala earthquakes clearly demonstrates that the origin of the long duration of strong ground shaking comes from the Sn and Lg wave trains. These S-wave arrivals are produced efficiently from shallow subduction earthquakes and are strongly enhanced during their propagation within the laterally heterogeneous waveguide produced by the subduction of the Cocos Plate beneath the Mexican mainland. The amplitude and duration of the Lg coda is also strongly reinforced by transmission through the Mexican Volcanic Belt from the amplification of S waves in the low-velocity surficial layer associated withS-to-P conversions in the volcanic zone. The further amplification of the large and long Lg wave train impinging on the shallow structure in the basin of Mexico City, with very soft soil underlain by nearly rigid bedrock with a strong impedance contrast, gives rise to the destructive strong ground shaking from the Mexican subduction earthquakes.

Regional Variation of

Records from broadband digital stations have allowed us to map regional variations of Lg coda Q across almost the entire United States. Using a stacked ratio method we obtained estimates of Q 0 (Lg coda Q at 1 Hz) and its frequency dependence, η, for 218 event-station pairs. Those sets of estimates were inverted using a back-projection method to obtain tomographic images showing regional variations of Q 0 and η. Q 0 is lowest (250–300) in the California coastal regions and the western part of the Basin and Range province, and highest (650–750) in the northern Appalachians and a portion of the Central Lowlands. Intermediate values occur in the Colorado Plateau (300–500), the Columbia Plateau (300–400), the Rocky Mountains (450–550), the Great Plains (500–650), the Gulf Coastal Plain and the southern portion of Atlantic Coastal Plain (400–500), and the portions of the Central Lowlands surrounding the high-Q region (500–550). The pattern of Q 0 variations suggests that the United States can be divided into two large Q provinces. One province spans the area from the Rocky Mountains to the Atlantic coast, is technically stable, and exhibits relatively high Q 0. The other extends westward from the approximate western margin of the Rocky Mountains to the Pacific coast, is tectonically active, and exhibits low Q 0. The transition from high to low Lg coda Q in the western United States lies further to the west than does an upper mantle transition for Q and electrical resistivity found in earlier studies. The difference in Q 0 between the western and eastern United States can be attributed to a greater amount of interstitial crustal fluids in the west. Regions of moderately reduced Q within the stable platform often occur where there are accumulations of Mesozoic and younger sediments. Reduced Q 0 in the southeastern United States may not be due to anelasticity but may rather be explained by a gradational velocity increase at the crust-mantle boundary that causes shear energy to leak into the mantle.

Lg

Regional seismograms were collected to image the lateral variations of Lg coda Q at 1 Hz (Q 0) and its frequency dependence (η) in the Middle East using a back-projection method. The data include 124 vertical-component traces recorded at 10 stations during the period 1986–1996. The resulting images reveal lateral variations in both Q 0 and η. In the Turkish and Iranian Plateaus, a highly deformed and tectonically active region, Q 0 ranges between about 150 and 300, with the lowest values occurring in western Anatolia where extremely high heat flow has been measured. The low Q 0 values found in this region agree with those found in other tectonically active regions of the world. Throughout most of the Arabian Peninsula, a relatively stable region, Q 0 varies between 350 and 450, being highest in the shield area and lowest in the eastern basins. All values are considerably lower than those found in most other stable regions. Low Q values throughout the Middle East may be caused by interstitial fluids that have migrated to the crust from the upper mantle, where they were probably generated by hydrothermal reactions at elevated temperatures known to occur there. Low Q 0 values (about 250) are also found in the Oman folded zone, a region with thick sedimentary deposits, η varies inversely with Q 0 throughout most of the Middle East, with lower values (0.4–0.5) in the Arabian Peninsula and higher values (0.6–0.8) in Iran and Turkey. Q 0 and η are both low in the Oman folded zone and western Anatolia.

Lg coda Q variation across Eurasia and its relation to crustal evolution

An extensive set of broadband and short‐period seismograms from regional seismic events has been used to derive a tomographic image of broad‐scale variations of Lg coda Q for most of Eurasia. The image indicates that Lg coda Q, at a frequency of 1 Hz, varies between about 200 and 1000. Lowest Q values (&lt;350) all lie within the Tethysides orogenic belt that was formed from the collision of the African/Arabian and Indian plates with the Eurasian plate. Relatively low Q values are found in the Arabian Peninsula (350–500), a region of recent uplift and volcanism, and in the eastern portion of the Altaid belt (∼450), a region covered by thick sediments. High Lg coda Q values (&gt;800) are found in four regions, the East European shield, the Siberian shield, the Indian shield, and the southwestern (and oldest) portion of the Altaid belt. The pattern of Lg coda Q variation across Eurasia suggests that crustal Q in any region is directly proportional to the length of time that has elapsed since the most recent episode of large‐scale compressional tectonic activity or upper mantle heating that occurred there. We interpret the low Q as resulting largely from hydrothermal fluids, generated by tectonic activity or heating, that now reside in cracks and permeable portions of the Eurasian crust. Crustal Q increases with time, after that activity, as fluids are lost to the surface or absorbed by retrograde metamorphism.

Lateral Variations of Lg Coda Q in Southern Mexico

—Broad band digital three-component data recorded at UNM, a GEOSCOPE station, were used to estimate Lg coda Q for 34 medium size (3.9 ≤m b ≤ 6.3) earthquakes with travel paths laying in different geological provinces of southern Mexico in an effort to establish the possible existence of geological structures acting as wave guides and/or travel paths of low attenuation between the Pacific coast and the Valley of Mexico. The stacked spectral ratio method proposed by XIE and NUTTLI (1988) was chosen for computing the coda Q. The variation range of Q 0 (Q at 1 Hz) and the frequency dependence parameter η estimates averaged on the frequency interval of 0.5 to 2 Hz for the regions and the three components considered are: i) Guerrero region 173 ≤Q¯ 0≤ 182 and 0.6 ≤Q¯ 0 ≤ 0.7, ii) Oaxaca region 183 ≤Q¯ 0 ≤ 198 and 0.6 ≤Q¯ 0 ≤ 0.8, iii) Michoacan-Jalisco region 187 ≤Q¯ 0 ≤ 204 and 0.7 ≤Q¯ 0 ≤ 0.8 and iv) eastern portion of the Transmexican Volcanic Belt (TMVB) 313 ≤Q 0≤ 335 and η = 0.9. ¶The results show a very high coda Q for the TMVB as compared to other regions of southern Mexico. This unexpected result is difficult to reconcile with the geophysical characteristics of the TMVB, e.g., low seismicity, high volcanic activity and high heat flow typical of a highly attenuating (low Q) region. Visual inspection of seismograms indicates that for earthquakes with seismic waves traveling along the TMVB, the amplitude decay of Lg coda is anomalously slow as compared to other earthquakes in southern Mexico. Thus, it seems that the high Q value found does not entirely reflect the attenuation characteristics of the TMVB but it is probably contaminated by a wave-guide effect. This phenomenon produces an enhancement in the time duration of the Lg wave trains travelling along this geological structure. This result is important to establish the role played by the transmission medium in the extremely long duration of ground motion observed during the September 19, 1985 Michoacan earthquake. ¶The overall spatial distribution of coda Q values indicates that events with focus in the Michoacan-Jalisco and Oaxaca regions yield slightly higher values than those from Guerrero. This feature is more pronounced for the horizontal component of coda Q. A slight dependence of average coda Q −1 on earthquake focal depth is observed in the frequency range of 0.2 to 1.0 Hz approximately on the horizontal component. Deeper (h > 50 km) events yield lower values of Q −1 than shallower events. For frequencies higher than 1.0 Hz no clear dependence of Q −1 on focal depth is observed. However, due to the estimates uncertainties this result is not clearly established.

Regional wave propagation in southern California and Nevada: Observations from a three‐component seismic array

We present results from a study conducted to characterize regional wave propagation in southern California and Nevada. The data are from events whose paths sample the southern Coast Ranges, the southern Sierra Nevada Mountains and the southern Basin and Range. These regional earthquake data were recorded by a small‐aperture three‐component seismic array located at the Piñon Flat Observatory, southern California. The data are analyzed using a statistical array processing algorithm that provides estimates of the propagation direction, frequency, wavenumber, and particle motion polarization characteristics of the coherent arrivals observed at the array. Perhaps the most striking feature shared by these data is the extent to which the P, S and Lg wave trains are composed of coherent, forward scattered/multipathed energy. The nature of the P, S and Lg wave trains suggests that spatially anisotropic scattering plays an important role in regional wave propagation in southern California and Nevada, and that scattering sources are located not on elliptical shells whose foci are the source and the receiver but instead in a limited volume subparallel to the path of the direct arrivals. The high‐amplitude, coherent nature of numerous of the Pg and Lg coda waves suggests that strong scattering in the source zone plays an important role in shaping the Pg and Lg wave trains. While the P, S and Lg wave trains for these regional events share some general characteristics they also exhibit fairly significant differences. These differences illustrate the important role regional and source zone structure play in shaping the wave field, and the need to account for both complex scattering and three‐dimensional propagation effects in regional waveform modeling. The coda for these regional data do not begin to exhibit a character consistent with the random scattering model for coda generation until the arrival of the surface waves. The fact that much of the surface wave coda cannot be effectively modeled as plane waves suggests that it is not the product of scattering from distant heterogeneities.

Excitation and propagation of Lg from earthquakes in central Asia with implications for explosion/earthquake discrimination

Broadband Lg records from 52 shallow earthquakes that occurred near the Balapan and Lop Nor nuclear test sites have been used to study spectral characteristics of the excitation and propagation of the Lg phase in central Asia. Using a source model that falls off as ω2 at high frequencies and a nonlinear inversion method, we simultaneously inverted for Lg seismic moments (M0), corner frequencies (fc), and path‐variable values of Q0 and η (QLg at 1 Hz and its power law frequency dependence). The logarithms of M0 obtained using Lg correlate linearly with logarithms of fc, the slope being about −3.56 when all data are used and −4.04 when data were sufficient that full inversions could be performed. These slopes differ only slightly from the slope of −3.83 obtained in our earlier studies using underground nuclear explosions. The fc values in this study are systematically lower than those found for explosions at the same M0 level. A comparison of corner frequencies and Lg moment values for unknown events with the plots for known events of this study therefore provides means to discriminate between earthquakes and explosions. The mb values are systematically lower for earthquakes of this study than for previously studied explosions at the same seismic moment; moreover, the logarithms of Lg moments obtained in this study correlate linearly, with a slope of 1.04, with mb values when all data are used and 1.21 when only well‐recorded events are used. These slopes are close to the slope of 1.19 obtained for explosions. The average QLg values along the paths used in this study agree well with those obtained using explosion data in central Asia and with reported Lg coda Q values.

Source spectra and site‐response estimates using the Coda of Lg waves in western Europe

Source spectra are investigated by using regional events recorded by the LDG French short‐period seismic network. Empirical relations describing the Lg coda envelope allow us to separate source, site and propagation effects. 28 events recorded by 14 stations of the LDG network are selected for this study. Source spectra and site‐response are determined in the frequency‐band of 1–16 Hz. A Lg‐coda magnitude can be defined from these source spectra. Magnitudes obtained by this technique are strongly correlated to local magnitudes given by the LDG bulletin but the standard deviation is reduced by a factor 3. Other parameters such as siteresponse, corner‐frequency and coda Q are determined very accurately.