No abstract

El objetivo de este trabajo es obtener secciones sísmicas para delimitar estructuras geológicas asociadas a yacimientos de hidrocarburos. Utilizamos datos de sísmica pasiva de un arreglo rectangular localizado al noreste de la República Mexicana y el método de Interferometría Sísmica (IS), el cual permite extraer la función de Green mediante el uso de correlaciones cruzadas (CC(t)) de trazas de ruido sísmico. Procesamos 6 líneas de 8 km de longitud aproximadamente, con 159 geófonos cada una. Las CC(t) obtenidas (también llamadas pseudo-secciones sísmicas de tiros virtuales) fueron procesadas para la extracción de información estructural y de velocidad subsuperficial, con los algoritmos que usualmente son utilizados en la industria de la sismología de exploración. Los resultados los reagrupamos en grupos de offset común (Common Offset Gathers, por sus siglas en ingles COG), para generar imágenes del campo de velocidades en el subsuelo en términos de secciones sísmicas de offset cero.  Realizamos un procesamiento que consistió en aplicar el método de la transformada de ondícula, para descomponer y reconstruir las CC(t) a partir de la ondícula de Meyer, mejorando la calidad de las CC(t). Posteriormente, generamos nuevas pseudo-secciones sísmicas considerando un offset en común y varias longitudes entre fuente-receptor. En este caso los resultados mejoraron substancialmente, mostrando los reflectores esperados. Observamos que, con un offset mayor, la profundidad de investigación aumentaba, aunque la resolución superficial disminuía. Finalmente, las secciones obtenidas fueron comparadas con una sección sísmica convencional (también llamadas secciones de sísmica “activa”) migrada en profundidad.

<p>In the framework of the international collaboration between Mexico and Europe for the development of geothermal energy (GEMex consortium), a seismic network of 45 seismic stations (25 broad-band and 20 short-period) was installed around the super-hot geothermal system of Los Humeros (Mexico) for more than one year. Los Humeros power plant is nested inside a quaternary caldera located in the eastern part of the Trans-Mexican Volcanic Belt that crosses the whole country from the Pacific coast to the Gulf of Mexico.</p><p>Among the several targets of the data collected by this network, an important task is to produce a seismic image of the caldera and of the geothermal reservoir. Here we present the 3D anisotropic shear wave velocity models retrieved by the seismic ambient noise tomography.</p><p>Thanks to the severe pre-processing of the whole seismic database we were able to obtain reliable and highly resolved models.</p><p>To carry out the model we applied a rigorous data quality assessment consisting in: 1) correction of the orientation of the sensors using the polarization of surface waves associated with tele-seismic and regional earthquakes, 2) assessment of the synchronization of the stations and correction of the times using daily cross-correlations functions, 3) finally to asses the quality of the stacked cross-correlations, knowed as Green’s functions (GF), we analyzed the noise sources directivity, inter-station distance and level of emergence of surface waves depending on the type of sensor used.</p><p>The processing allowed to pick clearly about 600 dispersion curves per velocity type (group and phase of R and L waves), using the NDCP code (Noisy Dispersion Curve Picking), that allows to display and select dispersion patterns both in time and frequency domain, for both causal and anti-causal part of the GF.</p><p>2D tomography maps were calculated from 0.5 to 9 s for each type of velocity. Depth inversion for the whole velocities types was carried out using surf96, allowing reconstructing the 3D anisotropic structure of the caldera for the first time.</p><p>The resulting models provides a larger view of the caldera and its anisotropic patterns down to 10 km depth. In these models, we were able to define the depth of the caldera rim, some important features of the internal part of the caldera and a low velocity body that could be associated with the hot sources feeding the reservoir. Our model are in strong agreement with those retrieved applying other geophysical methodologies (e.g. magnetotelluric, passive travel-time tomography, gravimetric, etc.).</p><p>This work is performed in the framework of the Mexican European consortium GeMex (Cooperation in Geothermal energy research Europe-Mexico, PT5.2 N: 267084 funded by CONACyT-SENER : S0019, 2015-04, and Horizon 2020, grant agreement No. 727550).</p>

<p>Joint inversion of surfaces and teleseismic converted waves is commonly used to retrieve seismic structures beneath a seismic station. Currently, this approach is routinely applied at global and regional scale to probe the structures of the mantle and the lower-crust. However, the difficulty to retrieve reliable converted waves at high frequencies (> 1 Hz) makes challenging to apply this technique to resolve structures at shallow depths (<20 km). Here we explore the feasibility of using a trans-dimensional Bayesian scheme based on a reversible jump Markov Chains Monte Carlo method, to resolve shallow structure at local scale. We use phase and group velocity dispersion curves for Love and Rayleigh waves, from 0.5 to 10 s and tele-seismic converted waves in a distance range from 30<sup>o</sup> to 95<sup>o</sup>. We explore the ability of different approaches to retrieve high frequency converted phases that will be used in the framework of the Bayesian inversion. We present preliminary tests of the reliability of the method and applications to experimental data collected in the super-hot geothermal field of Los Humeros, México. This work is performed in the framework of the Mexican European consortium GeMex (Cooperation in Geothermal energy research Europe-Mexico, PT5.2 N: 267084 funded by CONACyT-SENER: S0019, 2015-04, and Horizon 2020, grant agreement No. 727550).</p>

<p>In the framework of the GEMEX project (cooperation between Europe and Mexico for geothermal development), a dense network of 45 stations was deployed in 2017-2018 in the Los Humeros caldera, Mexico.</p><p>Thanks to this network an intense local seismic activity has been recorded in the geothermal field and surroundings, from which it has been possible to identify high-frequency Volcano-Tectonic events (VT, >10 Hz) and Long-Period events (LP, 1-8 Hz). The former set of events is mainly associated with the local tectonics and power plant activities; while the latter has been generally recorded after strong earthquakes (Mw>7) occurred in Mexico.</p><p>Consequently, we adapted and applied two tomographic techniques to generate highly resolved seismic models; 1) the Enhanced Seismic Tomography (EST) method using the travel times of local seismic events. The method incorporates the Double Difference tomography and the post-processing Weighted Average Method to generate Vp and Vs models, and 2) the surface wave tomography method based on ambient noise analysis. In this case, we generated 3D anisotropic models of phase and group velocities of the Rayleigh and Love waves from Green functions retrieved by cross-correlation of the continuous records.</p><p>Thanks to the severe pre-processing of the whole seismic database that allowed to correct several errors on the data, and to the methods applied, we were able to obtain reliable and highly resolved models with both techniques.</p><p>Finally, the two sets of events (VT and LP) have been relocated using the 3D seismic velocity models of the region in order to better characterize the structure of the geothermal field and identify regions where the fluids could have a role on the triggering of the LP seismic activity observed.</p><p>This work is performed in the framework of the Mexican European consortium GeMex (Cooperation in Geothermal energy research Europe-Mexico, PT5.2 N: 267084 funded by CONACyT-SENER: S0019, 2015-04, and Horizon 2020, grant agreement No. 727550).</p>