NS53A-01 INVITED
Detecting Fracture and Fault Zone Within an Unstable Mountain Slope in the Swiss Alps
Risks associated with sudden mountain-slope failures are escalating as a result of (i) rapidly expanding population centers, lifelines, and other critical infrastructure within mountain valleys and (ii) increases in exceptional climatic events and accelerated melting of alpine permafrost due to global warming. Accordingly, there is a need to implement suitable mitigation measures in the form of early warning systems and protective barriers. To design effective barriers, comprehensive knowledge of the locations and volumes of unstable rock is essential. To address these issues, we have been investigating an unstable mountain slope situated above the largest rockslide in recent Swiss history. We have further developed and applied a number of geophysical techniques at the study site. Innovative data processing schemes that included f-x-y deconvolution, topographic migration and semblance migration were required to obtain meaningful images of shallow- to steep-dipping fracture zones from 3-D surface ground-penetrating radar data. Critical additional details on the locations and geometries of the steep-dipping fracture zones were provided by single-hole radar data acquired in three moderately deep boreholes. Tomographic inversions of a comprehensive 3 D seismic refraction data set delineated a huge volume of very low quality crystalline rock with ultra-low to very low P-wave velocities of 500- 2700 m/s. These values were astonishingly low compared to the average horizontal P wave velocity of 5400 m/s determined from laboratory analyses of intact rocks collected at the site. The extremely low velocities extended to more than 35 m depth over a 200 x 150 m area that encompassed an actively moving segment of the mountain slope and a large part of the adjacent stationary slope. Finally, a specially designed microseismic network detected 223 microearthquakes during a 31-month monitoring period. These events, which had moment magnitudes of -2 to 0, were concentrated within 50-100 m of the surface in two zones, one that followed the recent rockslide scarp and one that spanned the volume of highest fracture zone/fault density.
NS53A-02
Reflection seismic imaging of a hydraulically conductive fracture zone in a high noise area, Forsmark, Sweden
High resolution reflection seismic methods have proven to be useful tools for locating fracture zones in crystalline rock. Siting of potential high-level nuclear waste repositories is a particularly important application of these methods. By using small explosive sources (15-75 grams), high resolution images of the sub-surface have been obtained in the depth range 100 m to 2 km in Sweden, Canada and elsewhere. Although ambient noise conditions in areas such as the Fennoscandian and Canadian shields are generally low, industrial noise can be high in some areas, particularly at potential sites suitable for repositories, since these are often close to existing infrastructure. In addition, the presence of this infrastructure limits the choice of sources available to the geophysicist. Forsmark, located about 140 km north of Stockholm, is one such potential site where reflection seismics have been carried out. Existing infrastructure includes nuclear reactors for power generation and a low- level waste repository. In the vicinity of the reactors, it was not possible to use an explosive source due to permitting restrictions. Instead, a VIBSIST system consisting of a tractor mounted hydraulic hammer was used in the vicinity of the reactors. By repeatedly hitting the pavement, without breaking it, at predefined sweeps and then stacking the signals, shot records comparable to explosive data could be generated. These shot records were then processed using standard methods to produce stacked sections along 3 profiles within the reactor area. Clear reflections are seen in the uppermost 600 m along 3 of these profiles. Correlation of crossing profiles shows that the strongest reflection (B8) is generated by a gently east-southeast dipping interface. Prior to construction of the reactors, several boreholes were drilled to investigate the bedrock in the area. One of these boreholes was located close to where two of the profiles cross. Projection of the B8 reflection into the borehole shows that the reflection corresponds to a hydraulically conductive fracture zone that was intersected at about 320 m depth.
NS53A-03
Ssismic Methodologies Applied To The Characterization Of Fractured Rock Massifs: Case Studies
The detailed characterization of fractured media in the shallow subsurface is becoming important. The detailed knowledge of the fracture network is mandatory in any hydrogeological model to constrain the potential pathways for water circulation. The geophysical methodolgies can provide a detailed image of the fractured rock and also the 3D distribution of physical properties. Two case studies are discussed in this work. The characterization of fractures in a waste disposal site and, the design and construction of a subway tunnel. In the first case, a multiseismic experiment was carried out in an old abandoned uranium mine. 2D and 3D seismic experiments including VSP, surface seismic reflection and travel time seismic tomography provided a 3D image of the internal structure of a granitic massif for hydrogeological studies of the preferred paths for the migration of contaminants. The tectonic stability of the site was also addressed by means of seismic measurements. The joint interpretation of all the available data enabled the interpretation of the low velocity anomalies in the 3D seismic tomography image as the fragile fractures and the alteration associated to them. A 3D image of the geometry of the heterogeneous weathered surface layer was also obtained. This surface is controlled by the complex network of faults and dykes observed in the area. The second case study involves 2D and 3D seismic experiments to aid the horizontal drilling of tunnels for a new subway line in Barcelona (Spain). Seismic data acquisition in a densely populated city is very difficult. The street layout determines the geometry of the acquisition experiments. The instrumentation can not always be located on the surface projection of the tunnel trace, therefore, pseudo 3D acquisition is required, deploying the instrumentation were it is possible. Furthermore, the shallow subsurface features extremely heterogeneous "weathered" layer of variable thickness (building fundations, sewage system, water supply conductions, etc), and the background noise is very high (car traffic, electricity lines, etc). Relatively old cities also lack a detailed geological control of the subsurface. Travel time seismic tomography provided a detailed characterization of the shallow subsurface including the 2D and 3D distribution of the physical properties and the reconstruction of the network of fractures (faults and dykes). The velocity models also imaged an important Miocene fault, quite different to the previous geological section derived from the core interpretation.
NS53A-04
Seismic Characterisation of Fractured Hard Rock, Drawing from Experience at the Olkiluoto Nuclear Waste Repository Site, Finland
The characterization of a deep-rock nuclear waste repository site requires an understanding of the structural geometry of the rockmass over a range of scales, from fractures identified in boreholes, to lower resolution, but wider coverage surface geological and geophysical data. Where complex fracture networks exist the inclusion of interpretations derived from VSP data, having an intermediate scale, leads to a robust structural model. VSP produces higher frequencies (up 200 Hz typically and up to 500 Hz occasionally) than surface seismic reflection and has the capability of estimating the 3D orientation of reflectors, including the sub-vertical. Seismic reflectors are typically assessed up to a 300-400 m lateral distance from the borehole and to depths exceeding 2000 m. The Olkiluoto rock mass consists of migmatised banded and foliated gneisses, cut by granite, amphibolite and metadiabase veins. The rock mass has undergone a polyphasic deformation, faulted and brecciated zones being related to the transition between the ductile and brittle deformation phases. VSP investigations were performed to determine the orientation and spatial continuity information of fracture zones and hydraulic conductive zones observed in boreholes. Fracture zone orientations were determined for typical trace lengths of 200 - 1000 metres, and a minimum thickness (set cut off limit) of two metres. Specialist VSP processing techniques, including 3- component Image Point Transform (IPT) (Cosma, 1995), polarization analysis and dip determination, were used to identify and map reflectors in 3D space. A verification procedure has been set to identify seismic reflectors in boreholes and attempt to correlate the geometrical and physical properties of the seismic reflecting features. Seismic features were found to be associated, not only with pronounced fracture zones, but also with lithological contacts. Consequently, reflectors without borehole control data may not directly be used as fracture zone indicators. However, once an identification procedure has been set, VSP has been found a useful tool for both structural and lithological modeling of crystalline rock. From a subset of 29 most proeminent features analyzed in a comprehensive interpretation exercise, 15 (50%) were clearly identified as fracturing-related and 14 (50%) were identified as combination of rock type/shearing and fracturing. Of the entire subset 24 (83%) were unambiguously explained. Without a firm explanation remained only one event and four of features although explained, had partly unconfirmed orientations. Boreholes are intersecting sub-horizontal features at optimal angles, making them better and more reliably represented in the site model. Sub-vertical events are somewhat more difficult to associate with the borehole data due to their less likely intersections with the boreholes, typically not being encountered in several boreholes.
NS53A-05 INVITED
Seismic Anisotropy of Soft Sands, Offshore Western AUstralia
Seismic anisotropy is commonly measured in sand shale environment. Intrinsic polar anisotropy of the shale and its effect on seismic data processing and analysis is well established and reasonably well understood. In sandstone, azimuthal anisotropy is often detected and is typically connected to an in situ stress regime and the brittleness of the rock. This type of anisotropy, commonly referred to as fractured induced anisotropy, has been widely and extensively studied as it directly affects both permeability and the strength of the rock. Hence fracture induced anisotropy is not only important for hydrocarbon exploration but also for geotechnical studies, underground mining, etc. Interestingly, in the last few years azimuthal anisotropy has also been detected in soft, poorly consolidated clean sands, mainly by cross-dipole sonic log measurements. This is somewhat surprising as in such soft, typically highly porous and permeable rocks stress induced fractures are unlikely to be abundant. In this study we analyse the anisotropy in such sand class using well-log measurements, three-component VSP data, as well as 2D and 3D surface seismic (reflection) data. High-quality cross-dipole sonic log measurements showed significant shear wave splitting over unconsolidated, highly porous and permeable sand interval. The shear wave anisotropy was computed to be around 10-15%. This is commonly seen as an indication that the rock is fractured and that the fractures are likely to be open. However, image log data over the same sand section suggested dilute most likely non-conductive fractures. Analysis of the shear wave splitting in VSP data also suggested low fracture density. The frequency content of the direct fast and slow shear waves on the VSP data was very similar, not supporting the presence of open fluid saturated fractures. Unfortunately, the evidence from the VSP data is not very compelling because the reservoir is thin compared to the wavelength and sampling interval of the VSP data. Further analysis of the soft sand anisotropy was conducted on surface seismic data. Magnitude of the overlain shale anisotropy was first established by measurements in the dominant horizontal stress direction. Subsequently pre-stack reflection amplitudes measured along several azimuths were matched to expected amplitudes from anisotropic AVO modelling. The results indicate that the anisotropy of the reservoir sands is high (more than 10%) at sonic frequencies but weak (about 2-3%), at seismic frequencies. We think this anisotropy is caused by the preferential closure of compliant inter-granular contacts oriented perpendicular to the principal horizontal stress. The effect is weaker at seismic frequencies since the wavelength in this case involves shales as well as sand. Furthermore, sonic anisotropy could have also been affected by the local stress conditions around the wellbore.
NS53A-06 INVITED
Radiomagnetotellurics in Near-Surface Studies of Azimuthal Electrical Anisotropy
Electrical and electromagnetic methods are often used to study azimuthal electrical anisotropy, characterized by different resistivities in the two principal orthogonal horizontal directions. A large fraction of these studies focus on determining the dominant directions of fracturing and in some cases even to estimate secondary porosity and possible anisotropy in hydraulic transmissivity. The azimuthal resistivity method has the following limitations: anisotropy must be suspected beforehand since a specialized electrode configuration is needed; the electrode configuration needs to be rotated with small increments at each sounding location; the soundings must be repeated with a small offset to distinguish anisotropy from heterogeneity. Consequently, the time-consuming character of such surveys results in only a few isolated soundings and an interpretation that is based on the inspection of the apparent resistivity data only. Tensor radio magnetotelluric (RMT) data do not suffer from these limitations: data collected over heterogeneous or homogeneous anisotropic structures can be distinguished by simple analysis of the measured transfer functions; no rotation of the measurement system is necessary in the field; tensor estimates obtained at each station can be inverted for a layered or smoothly varying electrical anisotropy model. As an illustration, we present tensor RMT data (12.7-243 kHz) collected along a 380-m long profile where limestones overlie shale. The data display a clear and consistent electrical anisotropy signature and were therefore inverted for a layered 1D model with azimuthal anisotropy. The resulting models indicate that the limestone formation has an anisotropy ratio of 12 and that the underlying shales are isotropic. The resulting models provide more regional and depth-dependent information than azimuthal electrical resistivity surveys presented in the literature.
NS53A-07
Anisotropy Characterization of Fractured Crystalline Bedrock Using Asymmetric Azimuthal Geoelectric Techniques
We examined the potential for geophysical characterization of fractured rock anisotropy by combining asymmetric configurations of azimuthal self potential (ASP) and azimuthal resistivity surveys (ARS), as previously demonstrated in the laboratory, at four field sites in the New Jersey Highlands (NJH) Province. There is a striking correlation between ASP measurements and fracture strike orientations at three of four sites investigated. ARS (electrical) data suggest three sites are overall heterogeneous and the fourth is anisotropic. The characteristic anisotropicity at the fourth site is controlled by a master structure; the NE-SW trending Lake Inez Fault Zone (LIFZ) that strikes at N10ºE and parallels the Wanaque River to the east side of the site. Inferred groundwater flow directions are comparable to the (1) positive polarity (+ve) and magnitude of site-specific SP, (2) local surface drainage, and (3) also conformable with the regional northwest and northeast fracture trend of the NJH. The ASP is ineffective at one heterogeneous site where there is a lack of correlation between ASP and fracture strike data, probably due to poor drainage where there are no distinct paths of flow defined along fractures. Quantitative analysis of the magnitude of the energy observed in the odd and even coefficients of the power spectra of self potential (SP) datasets analyzed using a Fourier series was useful for characterizing anisotropic or heterogeneous flow in the fracture network. For anisotropic flow, the odd coefficients (harmonics) were close to zero, whereas heterogeneous flow resulted in significant energy in the odd coefficients. The employment of asymmetric geoelectric arrays has allowed this quantitative distinction between anisotropy and heterogeneity in fractured bedrock. The results of our study suggest the ability to quantify hydraulic anisotropy with azimuthal self potential and the distinction between electrically-anisotropic and electrically-heterogeneous in the subsurface. These results represent a significant advancement over the use of traditional resistivity arrays in site characterization of fracture- dominated systems.