Fracture Traces Research Articles

Simplified 3D modelling method and its uncertainty analysis for adit fractures collected within short sampling windows

This study takes rock masses at the Songta Hydropower Station in SE China as examples to perform 3D discrete fracture network (DFN) modelling. Field fracture traces are collected from four adits (PD252, PD254, PD262 and PD264) with 1.6 m high sampling windows. The relationship between 3D fracture parameters and 2D collected trace information is investigated using data collected from short sampling windows. On the basis of this relationship, a simplified method is proposed to determine the 3D fracture density and disc diameter and generate 3D DFN. This method avoids the complex deduction processes of geometric fracture parameters, thereby simplifying 3D DFN modelling considerably. Although the method is developed on the basis of short-window collection, traditional data collected within a tall window are still applicable through the window dividing method. Short windows do not contain competent traces that reflect the structural information of rock masses. Therefore, errors are easily found during 3D DFN modelling. We propose a method to evaluate and reduce uncertainty effects. The uncertainties of fracture sets 1 and 2 are analysed by increasing the sampling size in a single structural domain, and the errors are acceptable.

A rigorous multiscale random field approach to generate large scale rough rock surfaces

Estimating the shear strength of large in situ rock discontinuities is non-trivial because of the multiscale nature of roughness and the fact that only a very limited extent of discontinuity morphology is visible along traces. Recently, a novel stochastic method based on random field theory and Monte Carlo semi-analytical estimation of shear strength was proposed. The method was validated at laboratory scale and its application to one large natural surface showed that it has the potential to bypass the scale effect. However, a critical issue was reported by the authors of the study: it was found that the random field approach used could not generate the correct distribution of gradients on the simulated surfaces, which translates into an inaccurate prediction of shear strength. The authors had to manually adjust the input of the model to achieve a satisfactory prediction. This paper presents a new multiscale approach using random field theory, which now allows a rigorous generation of large synthetic 3D surfaces with controlled distribution of asperity heights and gradients from the profile of a 2D fracture trace (as might be visible in a rock face) referred to as a seed trace. Each seed trace is first decomposed into three daughter profiles corresponding to three levels of roughness. The statistics of each daughter profile form the input of the random field model at each scale level, allowing synthetic daughter surfaces to be created at each scale level. The synthetic daughter surfaces are then superimposed to obtain a composite rough surface comprising three levels of roughness. The approach was successfully validated with 25 input seed traces, coming from 5 different natural surfaces. This rigorous multiscale approach is essential to apply the stochastic method that was recently developed to predict the shear strength of large in situ discontinuities.

Automated extraction and evaluation of fracture trace maps from rock tunnel face images via deep learning

This paper proposes an image-based method for automated rock fracture segmentation and fracture trace quantification. It is integrated using a CNN-based model named FraSegNet, a skeleton extraction algorithm, and a chain code-based polyline approximation algorithm. A rock tunnel fracture database with a total of 3,000 images of rock tunnel faces is established and selected to train and test the FraSegNet model. A comparison study is further conducted and shows that the FraSegNet model shows advanced performance in pixel-level fracture trace map extraction and noise reduction compared to other deep learning approaches and traditional image edge detection algorithms. Next, the skeletons of the predicted fracture trace maps are extracted and the corresponding polyline for each fracture skeleton is thus obtained and output as a text file composed of key nodes coordinates. The fracture trace characteristics (trace length, dip angle, density, and intensity) are acquired using node-based files. The quantitative evaluation of the proposed method illustrates that it can extract trace occurrences effectively and accurately. A case study of three full scale tunnel sections demonstrates the proposed method to be an efficient approach for acquiring and evaluating 2D fracture occurrences of under-construction rock tunnel faces.

The effect of strain-induced martensite transformation on strain partitioning and damage evolution in a duplex stainless steel with metastable austenite

Abstract Transformation induced plasticity (TRIP) assisted lean duplex stainless steels (LDSS) possess a multi-phase microstructure during deformation due to strain-induced martensite transformation (SIMT). The effect of SIMT on the strain partitioning into constituent phases, which in turn affects the SIMT kinetics is highlighted. The individual stress-strain relationships of each phase are obtained via a microstructure-based model and the gradual fragmentation of austenite grain resulted from SIMT is especially considered. A modified model of partitioned strain in the austenite considering the contribution of martensite is proposed, whereby the actual SIMT kinetics in the parent austenite phase can be quantitatively determined. Besides, the contributed strain of each constituent phase during the progress of deformation can be evaluated directly. Furthermore, EBSD and in-situ tensile tests were carried out to characterize the influences of SIMT and resulting high strain localization on the damage evolution. The cracking nucleation initiates at the α/α′ interface as well as the inside of ferrite near the interface. Subsequently, they grow along the α/α’ interface with further straining and, finally, merge into one interfacial crack. The combined crack forks off at the interface with small curvature, then the secondary cracks penetrate into austenite/martensite layer and ferrite layer, respectively, resulting in the final failure. The overall fracture of LDSS shows the ductile feature, although the ferrite and the martensite in the local region of fracture surface show the trace of quasi-cleavage fracture, the remained austenite exhibits dimple-typed fracture.

Effects of fracture parameters and roughness on heat‐flow coupling in rock masses with two‐dimensional fracture networks

AbstractDiscrete fracture network (DFN) is an effective means of describing the coupling of heat flow in an underground fractured rock mass. In this paper, an improved DFN is proposed by introducing the correlation function of fracture trace length and aperture, which is more consistent with the real fracture data. Next, based on the improved model, the influence of fracture roughness on the fluid flow and heat transmission was evaluated, and the relationship between the fracture aperture and the joint roughness coefficient (JRC) is established. Finally, based on the exponential function between confining pressure and aperture, the influence of confining pressure on the heat‐flow coupling process is considered in the improved model. Besides, the reliability of the model was verified by comparing with the analytical solution of the two‐dimensional single‐fracture heat‐flow coupling problem. The results show that under the same conditions, considering the correlation between the geometric parameters of the fracture, the seepage and heat transfer rates of the model increased, the outlet boundary flow reached the maximum value, and the average outlet temperature dropped rapidly. However, the fracture roughness reduces the outlet flow rate and the decline rate of average temperature. The confining pressure will lead to a decrease of about 3.5% in the outlet flow of the model, which is consistent with the seepage law in practical engineering. The model presented in this paper is an effective supplement to the two‐dimensional fracture network heat‐flow coupling model and can provide a theoretical basis and a numerical calculation tool for related underground rock mass engineering.

Feasibility of three-dimensional CT reconstruction in assessing mandibular fractures

Aim: Evaluation of the reliability of 3D computed tomography (3D-CT) in the diagnosis of mandibular fractures. Methods: A cross-sectional, quantitative and qualitative study was carried out, through the application of a questionnaire for 70 professionals in the area of Oral and Maxillofacial Surgery and Radiology. 3D-CT images of mandibular fractures were delivered to the interviewees along with a questionnaire. Participants answered about the number of traces, the region and the type of fracture. The correct diagnosis, that is, the expected answer, was based on the reports of a specialist in oral and maxillofacial radiology after viewing the images in the axial, sagittal and coronal sections. The resulting data from the interviewees was compared with the expected answer and then, the data was analyzed statistically. Results: In the sample 56.9% were between 22 and 30 years old, 52.8% were oral and maxillofacial surgeons (OMF), 34.7% were residents in OMF surgery and 12.5% OMF radiologists. Each professional answered 15 questions (related to five patients) and 50.8% of the total of these was answered correctly. Specialists in Oral and Maxillofacial Surgery and Traumatology correctly answered 53.9%. Interviewees with experience between 6 and 10 years correctly answered 58.2%. In identifying fracture traces, 46.1% of the questions were answered correctly. In terms of location, 5.6% of interviewees answered wrongly while 14.2% answered wrongly regarding classification. Conclusion: 3D computed tomography did not prove to be a reliable image for diagnosing mandibular fractures when used alone. This made necessary an association with axial, sagittal and coronal tomographic sections.

Clinical symptoms of iodine and zinc deficiency in dairy cows

Numerous clinical observations indicate that the silage-concentrate type of dairy cows feeding causes a cascade of pathobiochemical processes occurring by the type of metabolic acidosis, ketosis, hepatosis, with a violation of protein, fat and carbohydrate metabolism. Animals cannot realize the genetically inherent productivity potential.The culling of cows from the main herd is more than 37%. The main reasons for culling cows are hypo - and agalactia as a result of mastitis and ketosis, as well as infertility against the background of repeated infertile inseminations. Allotriophagy, unsteadiness of teeth, resorption of the last tail vertebrae, curvature of the vertebral column, traces of rib fractures, difficulty in lifting and moving when the content of total calcium and inorganic phosphorus in the blood is at the lower limit of physiological values indicate osteodystrophy. Exo-or endophthalmos, myxedema, forelock and mane, disheveled and curly hair, abortions, the birth of premature, ugly, stillborn, sparsely haired or “naked” calves against the background of low levels of protein bound iodine (PBI) in the blood indicate iodine deficiency in their body (hypomicroelementosis J). Coarsening and increased folding of the skin in the area of the head, neck and withers, active keratization on the skin in the area of articular surfaces and croup, thickening of the corolla zone and the border of the hooves in the form of a roller, ski-like growth and deformation of the hoof horn, enlargement and deformation of the joints, decreased reproductive function and low levels of zinc in the blood indicate parakeratosis of cattle (zinc deficiency).

Multidisciplinary Approach to the Treatment of an Unusual Crown Fracture after Trauma: Alternative Technique Proposal

This case report presents multidisciplinary management of a sub gingival crown-root fracture of a 15-year-old patient referred to the dental trauma clinic with an extensive tooth fracture of the permanent maxillary central incisors after a traumatic injury during a volleyball game. This work aimed to present an alternative approach for complicated dental fracture after trauma obtaining immediate resolution. A unique aspect of the case was that the patient was referred from the semiology clinic without pain and with signs of extensive cracking after 13 days of trauma. However, no endodontic involvement was observed either clinically or radio graphically. After examining a probe, the fracture trace dislocated with pulp exposure and sub-gingival extension. An alternative isolation technique was presented to enable immediate endodontic treatment and restoration of fractured teeth.

Fracture Mechanism in Overlying Strata during Longwall Mining

We used the key stratum theory to establish a more realistic thin‐plate mechanical model of elastic foundation clamped boundary and study the fracture mechanism of overlying strata during longwall mining. We analyzed the fracture characteristics and factors affecting fracture of the key stratum combined with the Mohr–Coulomb yield criterion. Besides, we used numerical simulation methods to verify the evolution pattern of the overlying strata fracture. The results show that the fracture mechanisms of the elastic foundation clamped structure’s key stratum varied depending on the position under longwall mining. The advanced coal wall area of the upper surface is a compressive‐shear fracture. The center area of the lower surface is a tensile fracture. With the increase of the excavation length and the load of the key stratum, the central area and the advanced coal wall area of the long side are fractured before the advanced coal wall area of the short side. With the increase of flexural rigidity of the key stratum, the advanced coal wall area of the long side fractures before the central area and the advanced coal wall area of the short side. With the increase of the foundation modulus and the advanced load of the key stratum, the central area fractures before the surrounding advanced coal wall area. The advanced influence distance was positively correlated with the key stratum’s flexural rigidity and advanced load and negatively correlated with the foundation modulus and excavation length. The advanced influence distance was not affected by the load of the key stratum. The numerical simulation results show that, with the increase of the mining area, the fracture trace of overlying strata in goaf extended to the coal wall’s interior. The fracture range of overlying strata is larger than that of the miningd: area. This study has a practical value for water disasters, gas outbursts, and rock strata control.

Three-Dimensional Separation and Characterization of Fractures in X-Ray Computed Tomographic Images of Rocks

Open fractures can affect petrophysical properties of their host rock masses, as well as fluid transport and storage, so characterization of them is important to both industrial and research scientists. X-ray Computed Tomography (CT), a non-destructive technique for 3D imaging of various materials, shows such fractures well in rock samples. However, separation and characterization of fractures in CT data is complicated when a scanned sample contains narrow and intersecting fractures, because narrow fractures become blurred when thinner than the scanner resolution and their value approximates the one of the matrix, and because intersecting features are difficult to individually characterize. In this paper, we present a new approach for an objective and efficient characterization of the fracture network inside CT scans of rock samples. We have developed algorithms, implemented as Python scripts, that measure fracture aperture-related parameters, and that separate connected fractures and fracture intersections within CT images of the sample. The CT images are composed of stacks of 2D images in the plane parallel to X-Y (equally spaced), where each pixel has a value related to the attenuation of the X-rays within the materials that make up the core at that location and is generally displayed using a gray-scale colormap. As the gray values in the reconstructed images drop within fractures, our algorithm is able to identify such drops and record the lowest gray value in every drop as a Fracture Trace Point (FTP). For every FTP, parameters related to the local fracture width and the three-dimensional orientation of the FTPs surrounding it are measured. A second step involves the separation of individual fractures and their intersections points. This allows information about a number of FTP measurements on the same fracture (or intersections) to be combined to characterize that feature. We demonstrate that our methods quantify fractures and their intersections in high detail through analysis of an experimentally-deformed granite sample, within which we characterize fracture size, orientation, and intensity. The methodologies can be also used to characterize sub-planar features in other types of datasets. Python implementations of our algorithms are freely available on GitHub repositories.

Fracture behavior of Longmaxi shale with implications for subsurface applications

Fracture mechanical properties of shales have gained continued interest recently due to their critical role in many shale-related applications such as unconventional petroleum systems and subsurface carbon and waste disposal. However, due to their strong reactive nature to fluids, fracture mechanical properties of shales have not been extensively studied like other rock types. A more comprehensive understanding of fracture system development in shales under reactive fluids is needed for subsurface applications. We have measured fracture mechanical properties of Longmaxi shale outcrop samples in the Sichuan Basin, China, under different environments of ambient air, deionized water, saline fluids of NaCl and KCl at 0.5 M concentration, and acidic HCl fluid at pH of 5. All aqueous fluids tested show strong weakening effects on fracture propagation compared to the air environment, with the fracture toughness reduced by 75% and the subcritical fracture growth index reduced by 50%. Microstructural analysis reveals the predominantly grain boundary opening cracking mode for all tests, but the fracture traces branch more in reactive aqueous fluids. Natural fractures are comparable with artificial fractures in morphology. Bedding-perpendicular opening mode fractures with multiple-stage fracture fillings of calcite, quartz, and organic matters develop well in natural fractures. Our results suggest that clay mineral hydration and expansion are the main cause for the fluid weakening effect in Longmaxi shale, which has substantial implications for subsurface shale failure processes.

Effect of long-term storage on the electronic structure of semiconducting silicon wafers implanted by rhenium ions

Rhenium-doped silicon samples had been synthesized by means of re-pulsed ion implantation (fluences 2.5 × 1017 cm−2, 1 × 1017 cm−2 and 5 × 1016 cm−2) and then stored under natural air-media conditions 5 years. The electronic structure of these samples was studied using combined XPS-and-DFT method. It was confidently established and demonstrated by means of XPS electronic structure mapping (core levels, valence band) that weak acidification occurs. The influence of air media leads to O…Si–O…[Re0] and …Re–Si–O… bonding appearance in the structure of 5-year aged samples rather than to normal rhenium oxidation. No any traces of RexOy sub-oxide fractures were neither found experimentally nor predicted theoretically. DFT-derived acidification scenario points out energetically preferable Re atoms pairing when low concentrations ion-doping regime is applied and metallic Re clusterization in the high concentrations ion-doping regime. Experimentally established essential transformation of initial vicinity states in the top of valence bands by means of Re 5d doping is in a good agreement with DFT calculations of electronic structure and formation energy values of acidification process.

Mapping the fracture network in the Lilstock pavement, Bristol Channel, UK: manual versus automatic

Abstract. The 100 000 m2 wave-cut pavement in the Bristol Channel near Lilstock, UK, is a world-class outcrop, perfectly exposing a very large fracture network in several thin limestone layers. We present an analysis based on manual interpretation of fracture generations in selected domains and compare it with automated fracture tracing. Our dataset of high-resolution aerial photographs of the complete outcrop was acquired by an unmanned aerial vehicle, using a survey altitude optimized to resolve all fractures. We map fractures and identify fracture generations based on abutting and overprinting criteria, and we present the fracture networks of five selected representative domains. Each domain is also mapped automatically using ridge detection based on the complex shearlet transform method. The automatic fracture detection technique provides results close to the manually traced fracture networks in shorter time but with a bias towards closely spaced Y over X nodes. The assignment of fractures into generations cannot yet be done automatically, because the fracture traces extracted by the automatic method are segmented at the nodes, unlike the manual interpretation in which fractures are traced as a path from fracture tip to fracture tip and consist of several connected segments. This segmentation makes an interpretation of relative age impossible, because the identification of correct abutting relationships requires the investigation of the complete fracture trace by following a clearly defined set of rules. Generations 1 and 2 are long fractures that traverse all domains. Generation 3 is only present in the southwestern domains. Generation 4 follows an ENE–WSW striking trend, is suborthogonal to generations 1 and 2, and abuts on them and generation 3, if present. Generations 5 is the youngest fracture set with a range of orientations, creating polygonal patterns by abutting at all other fracture generations. Our mapping results show that the northeastern domains only contain four fracture generations; thus, the five generations of the outcrop identified in the southwestern domains are either not all present in each of the five domains or vary locally in their geometry, preventing the interpreter from linking the fractures to their respective generation over several spatially separate mapping domains. Fracture intensities differ between domains where the lowest is in the NE with 7.3 m−1 and the highest is in the SW with 10 m−1, coinciding with different fracture orientations and distributions of abutting relationships. Each domain has slightly different fracture network characteristics, and greater connectivity occurs where the development of later shorter fractures is not affected by the stress shadowing of pre-existing longer fractures.

Data acquisition by digitizing 2-D fracture networks and topographic lineaments in geographic information systems: further development and applications

Abstract. Understanding the impact of fracture networks on rock mass properties is an essential part of a wide range of applications in geosciences from understanding permeability of groundwater aquifers and hydrocarbon reservoirs to erodibility properties and slope stability of rock masses for geotechnical engineering. However, gathering high-quality, oriented-fracture datasets in the field can be difficult and time-consuming, for example, due to constraints on field work time or access (e.g. cliffs). Therefore, a method for obtaining accurate, quantitative fracture data from photographs is a significant benefit. In this paper we describe a method for generating a series of digital fracture traces in a geographic information system (GIS) environment, in which spatial analysis of a fracture network can be carried out. The method is not meant to replace the gathering of data in the field but to be used in conjunction with it, and it is well suited when field work time is limited or when the section cannot be accessed directly. The basis of the method is the generation of the vector dataset (shapefile) of a fracture network from a georeferenced photograph of an outcrop in a GIS environment. From that shapefile, key parameters such as fracture density and orientation can be calculated. Furthermore, in the GIS environment more complex spatial calculations and graphical plots can be carried out such as heat maps of fracture density. Advantages and limitations compared to other fracture network capture methods are discussed.

GENERATIVE ADVERSARIAL NETWORKS AS A NOVEL APPROACH FOR TECTONIC FAULT AND FRACTURE EXTRACTION IN HIGH-RESOLUTION SATELLITE AND AIRBORNE OPTICAL IMAGES

Abstract. We develop a novel method based on Deep Convolutional Networks (DCN) to automate the identification and mapping of fracture and fault traces in optical images. The method employs two DCNs in a two players game: a first network, called Generator, learns to segment images to make them resembling the ground truth; a second network, called Discriminator, measures the differences between the ground truth image and each segmented image and sends its score feedback to the Generator; based on these scores, the Generator improves its segmentation progressively. As we condition both networks to the ground truth images, the method is called Conditional Generative Adversarial Network (CGAN). We propose a new loss function for both the Generator and the Discriminator networks, to improve their accuracy. Using two criteria and a manually annotated optical image, we compare the generalization performance of the proposed method to that of a classical DCN architecture, U-net. The comparison demonstrates the suitability of the proposed CGAN architecture. Further work is however needed to improve its efficiency.

Coupling effect of power-law fluid properties and scaled fractal characteristics on flow through fractured media

Non-Newtonian fluid flow through fractured media widely exists in both natural and man-made materials, where fractures have been shown to have self-similar and fractal characteristics. In this study, a new analytical approach is developed to calculate the flowrate of a power-law fluid through fractal fractures, in which fracture trace length follows a fractal distribution and the aperture and trace length are related by a scaled power-law function. The average fracture flowrate in the fractal fractures is obtained through integrating flows in all fractures. The interplay of the power-law properties and fractal characteristic parameters on the flow through the fractal fractures is quantified and discussed. The results demonstrate that the average flowrate decreases with the fractal dimension of fracture length and scaling exponent. The increase in the trace length range ratio leads to the increased flowrate. The power-law index plays a significant role in the flowrate through fractal fractures coupling with the effect of pressure head gradient. The increase in pressure head gradient amplifies the effect of power-law fluid index on the flowrate in fractal fractures.

An Investigation of the Factors Controlling the Mechanical Behaviour of Slender Naturally Fractured Pillars

Previous studies on the strength and deformability of pillars in hard-rock mines have demonstrated that the influence of natural fractures diminishes with increasing pillar width. Jointing conditions and fracture intensity/intersections play a critical role in the behaviour of slender pillars (width-to-height ratio less than 0.5). An important challenge is the characterization of rock mass damage as a function of the natural fracture network and loading conditions. In this paper, we present advanced pre-processing methodologies to measure fracture intensity and fracture density prior to loading and relate those measures to the evolution of damage (intensity and density) with increasing pillar load. A hybrid finite-discrete element method (FDEM) with fracture mechanics capabilities is used to simulate the mechanical behaviour of synthetic pillar models with different embedded discrete fracture traces of similar total length. Newly developed post-processing methodologies are used to capture the location of damage events, track rock mass damage intensity and density as the simulation progresses and relate those changes to the modelled pillar strength. We believe the results of this study provide useful insights on rock mass damage evolution.

Reliability analysis of seismic attribute in the detection of fault-karst

In the Tarim Basin, various irregular fractured-vuggy reservoirs have developed along with the main faults. These reservoirs are geologically defined as carbonate fault karst. In the past few years, seismic attributes have been widely used for the identification and evaluation of fault karst. However, there has been less reliability analysis regarding their usage. Imaging using the theoretical fault-karst velocity model can reflect the shapes and distributions of fractures and vugs, whereas imaging using the background velocity can simulate seismic data in real cases. We have adopted an approach based on typical fault-karst theoretical forward modeling to evaluate the reliability of seismic attributes in practical applications. First, we extract various attributes from the images using the theoretical velocity and the background velocity using similarity estimation between them to optimize the sensitive attributes. The analysis result indicates that the instantaneous phase, variance, amplitude gradient, coherence, and texture entropy are more suitable to characterize the anomalies of fractures and vugs with prediction accuracy of 71.7%. Because fracture orientation and density are the key parameters for quantifying the differences between the two images, taking coherence as an example, we extract the fracture traces through circular scanlines and circular windows based on the optimized attributes. The coincidence rate between the predicted fracture density and the known model reaches 83%, and that between the predicted fracture orientation and the known model is greater than 95%. With this remarkable coincidence, we can conclude that optimized seismic attributes are reliable for characterizing fractured-vuggy reservoirs.

Effect of Density, Trace Length, Aperture, and Direction Angle on Permeability Performance of Fracture Networks

A fractured rock mass was characterized by strong heterogeneity, discontinuity, and anisotropy. The study of flow in a fractured rock mass was a complex and challenging task. As the main seepage channels of a fractured rock mass, the fracture network was mainly affected by the fracture density, trace length, aperture, and direction angle. In this paper, the influence of fracture parameters on the permeability of the fracture network was studied by orthogonal experimental design and numerical modeling. A series of fracture networks were established to obtain seepage rules and permeability. The results showed that the increase in fracture density, trace length, and aperture increased the permeability of the model and that the fracture direction angle affected the pressure gradient and the anisotropy of the seepage in the model. The sensitivity of effective parameters on the discrete fracture network permeability performance was obtained. The influence factors of K1 were density, trace length, aperture, and direction angle in that order. The influence factors of K2 were direction angle, density, aperture, and trace length, in that order. The influence factors of K1/K2 were direction angle and density, aperture, and trace length in that order. These results could be used for parameter adjustment of fracture network modeling and determining the complex seepage characteristics in naturally fractured reservoirs.

Which fractures are imaged with Ground Penetrating Radar? Results from an experiment in the Äspö Hardrock Laboratory, Sweden

Identifying fractures in the subsurface is crucial for many geomechanical and hydrogeological applications. Here, we assess the ability of the Ground Penetrating Radar (GPR) method to image open fractures with sub-mm apertures in the context of future deep disposal of radioactive waste. GPR experiments were conducted in a tunnel located 410 m below sea level within the Äspö Hard Rock Laboratory (Sweden) using 3-D surface-based acquisitions (3.4 m × 19 m) with 160 MHz, 450 MHz and 750 MHz antennas. The nature of 17 identified GPR reflections was analyzed by means of three new boreholes (BH1-BH3; 9–9.5 m deep). Out of 21 injection and outflow tests in packed-off 1-m sections, only five provided responses above the detection threshold with the maximum transmissivity reaching 7.0 × 10−10 m2/s. Most GPR reflections are situated in these permeable regions and their characteristics agree well with core and Optical Televiewer data. A 3-D statistical fracture model deduced from fracture traces on neighboring tunnel walls show that the GPR data mainly identify fractures with dips between 0 and 25°. Since the GPR data are mostly sensitive to open fractures, we deduce that the surface GPR method can identify 80% of open sub-horizontal fractures. We also find that the scaling of GPR fractures in the range of 1–10 m2 agrees well with the statistical model distribution indicating that fracture lengths are preserved by the GPR imaging (no measurement bias). Our results suggests that surface-GPR carries the resolution needed to identify the most permeable sub-horizontal fractures even in very low-permeability formations, thereby, suggesting that surface-GPR could play an important role in geotechnical workflows, for instance, for industrial-scale siting of waste canisters below tunnel floors in nuclear waste repositories.