Tomographic inversion

In this study we focused on seismic attenuation (1/ Q c ) tomography in Northern Morocco. For this purpose, two different models are employed: The Single Backscattering model hypothesis of to calculate values of Coda Q (Q c ) and the Back-projection technique of to estimate lateral variation in Q c via a tomographic inversion. For this investigation, the Coda Q method is applied to a number of 94 local earthquakes with a magnitude between Ml=0.7 and Ml=4. The digital seismograms of these earthquakes were recorded during the year 2008 by both local temporary and permanent broadband seismic station network deployed in Northern of Morocco. The Q c quality factor values have been computed at central frequencies 0.75, 1.5, 3, 6 and 12 Hz. The lapse time windows are restricted to 30s in order to sample the earth's crust only. The Q c results indicate that strong frequency dependence follow a power law for the entire area. The preliminary results of seismic Coda Q c attenuation tomography shows a dependence at each frequency band, between seismic attenuation and the geology structure units in the study area, especially in the region of Al Hoceima and the eastern part of the Rif which are characterized by high attenuation values due to active faults area, while low attenuation values are seen in the west and the south of the Rif in high frequencies.

In this study we focused on seismic attenuation (1/ Q c) tomography in Northern Morocco. For this purpose, two different models are employed: The Single Backscattering model hypothesis of Aki and Chouet (1975) to calculate values of Coda Q (Q c) and the Back-projection technique of Xie and Mitchell (1990) to estimate lateral variation in Q c via a tomographic inversion. For this investigation, the Coda Q method is applied to a number of 94 local earthquakes with a magnitude between Ml=0.7 and Ml=4. The digital seismograms of these earthquakes were recorded during the year 2008 by both local temporary and permanent broadband seismic station network deployed in Northern of Morocco. The Q c quality factor values have been computed at central frequencies 0.75, 1.5, 3, 6 and 12 Hz. The lapse time windows are restricted to 30s in order to sample the earth's crust only. The Q c results indicate that strong frequency dependence follow a power law for the entire area. The preliminary results of seismic Coda Q c attenuation tomography shows a dependence at each frequency band, between seismic attenuation and the geology structure units in the study area, especially in the region of Al Hoceima and the eastern part of the Rif which are characterized by high attenuation values due to active faults area, while low attenuation values are seen in the west and the south of the Rif in high frequencies.

This paper presents the application of a probabilistic approach for variational inversion in acoustic tomography. The aim is to determine the time-evolving, rangeaveraged, vertical profile of speed of sound) , ( z c in a shallow water environment from the acoustic pressure fields generated by a monochromatic sound source and measured on a sparse vertical hydrophone array. A variational approach that minimizes a cost function which measures the distance between observations and their modelled counterparts is used. As the tomographic inversion is an ill-posed problem a regularization term in the form of a quadratic restoring term to a background is added. To avoid inverting for the variancecovariance matrix associated with the above weighted quadratic background, it is proposed to model the sound speed vector using probabilistic principal component analysis (PPCA). The probabilistic PCA introduces an optimum reduced number of non correlated latent variables  which determine a new control vector and introduce a new regularization term, expressed as   T. PPCA represents a rigorous formalism for the use of a priori information and allows for an efficient implementation of the variational inverse method. In the present work the probabilistic PCA is applied to an acoustic tomography scenario in the South Elba environment.

As part of a small scale sequestration test (about 1500 tonsof CO2) in a saline aquifer, time-lapse borehole seismic surveys wereconducted to aid in characterization of subsurface CO2 distribution andmaterial property changes induced by the injected CO2. A VSP surveydemonstrated a large increase (about 75 percent) in seismic reflectivitydue to CO2 injection and allowed estimation of the spatial extent of CO2induced changes. A crosswell survey imaged a large seismic velocitydecrease (up to 500 m/s) within the injection interval and provided ahigh resolution image of this velocity change which maps the subsurfacedistribution of CO2 between two wells. Numerical modeling of the seismicresponse uses the crosswell measurements to show that this small CO2volume causes a large response in the seismic reflectivity. This resultdemonstrates that seismic detection of small CO2 volumes in salineaquifers is feasible and realistic.

In this study we focused on seismic attenuation (1/ Q c) tomography in Northern Morocco. For this purpose, two different models are employed: The Single Backscattering model hypothesis of Aki and Chouet (1975) to calculate values of Coda Q (Q c) and the Back-projection technique of Xie and Mitchell (1990) to estimate lateral variation in Q c via a tomographic inversion. For this investigation, the Coda Q method is applied to a number of 94 local earthquakes with a magnitude between Ml=0.7 and Ml=4. The digital seismograms of these earthquakes were recorded during the year 2008 by both local temporary and permanent broadband seismic station network deployed in Northern of Morocco. The Q c quality factor values have been computed at central frequencies 0.75, 1.5, 3, 6 and 12 Hz. The lapse time windows are restricted to 30s in order to sample the earth's crust only. The Q c results indicate that strong frequency dependence follow a power law for the entire area. The preliminary results of seismic Coda Q c attenuation tomography shows a dependence at each frequency band, between seismic attenuation and the geology structure units in the study area, especially in the region of Al Hoceima and the eastern part of the Rif which are characterized by high attenuation values due to active faults area, while low attenuation values are seen in the west and the south of the Rif in high frequencies.

In this study we focused on seismic attenuation (1/ Q c) tomography in Northern Morocco. For this purpose, two different models are employed: The Single Backscattering model hypothesis of Aki and Chouet (1975) to calculate values of Coda Q (Q c) and the Back-projection technique of Xie and Mitchell (1990) to estimate lateral variation in Q c via a tomographic inversion. For this investigation, the Coda Q method is applied to a number of 94 local earthquakes with a magnitude between Ml=0.7 and Ml=4. The digital seismograms of these earthquakes were recorded during the year 2008 by both local temporary and permanent broadband seismic station network deployed in Northern of Morocco. The Q c quality factor values have been computed at central frequencies 0.75, 1.5, 3, 6 and 12 Hz. The lapse time windows are restricted to 30s in order to sample the earth's crust only. The Q c results indicate that strong frequency dependence follow a power law for the entire area. The preliminary results of seismic Coda Q c attenuation tomography shows a dependence at each frequency band, between seismic attenuation and the geology structure units in the study area, especially in the region of Al Hoceima and the eastern part of the Rif which are characterized by high attenuation values due to active faults area, while low attenuation values are seen in the west and the south of the Rif in high frequencies.

Como requisito previo para la consecución de imágenes tomográficas se estudian tres modelos de Moho. Para la región de Ecuador se analizan el modelo gravimétrico de Chambat-Valette, un modelo de velocidades de propagación de ondas sísmicas en tres dimensiones y un modelo de observaciones gravimétricas satelitales. Se presenta como resultado un modelo que junta las mejores características de estas tres investigaciones previas.

Como requisito previo para la consecución de imágenes tomográficas se estudian tres modelos de Moho. Para la región de Ecuador se analizan el modelo gravimétrico de Chambat-Valette, un modelo de velocidades de propagación de ondas sísmicas en tres dimensiones y un modelo de observaciones gravimétricas satelitales. Se presenta como resultado un modelo que junta las mejores características de estas tres investigaciones previas.

We investigate various issues related to the propagation of seismic waves in Eurasia during the period August 16, 1990 – December 15, 1992 under the Air Force Phillips Laboratory contract F19628-90-C-0158. This final report covers the research accomplished under the 2 tasks that have not been reported previously in the first annual report PL-TR-92-2042.  Section 1 describes how we derived a spherically symmetric velocity structure for the Garm region in Central Asia with travel-time residuals of regional phases recorded at the IRIS station GAR and the CDSN station WMQ. In Section II, we conducted a tomographic inversion to determine the regional Lg Q structure in the Iranian Plateau. Although the whole of Iran can be regarded as a region of very low Q, applying a simple averaged attenuation coefficient (Q) for the whole plateau would be inappropriate for the Lg magnitude determination. The regional variation of the anelastic attenuation parameter is significant enough that it must be taken into account in calibrating each monitoring station for a reliable magnitude scale in monitoring possible clandestine tests from a vast area. Section III of this report gives a perspective overview of the whole project.