Fractured crystalline rocks form highly heterogeneous subsurface systems in which fluid flow is controlled by the connectivity and aperture of fracture networks. These hydraulic properties are controlled by in-situ stress fields, which, in the shallow crust, results from the combined effects of regional tectonic stress, gravitational forces, and topographic relief. Although stress-dependent fracture behavior has been extensively studied at the fracture scale, the role of spatially variable stress tensors in controlling hydraulic conductivity at the catchment scale remains poorly constrained.In this study, we investigate how topography and tectonic stresses interact with fracture network geometry to produce heterogeneity and anisotropy in hydraulic conductivity across a high mountain catchment. A three-dimensional geomechanical model is used to compute spatially variable stress tensors for both synthetic and real topographic surfaces. The resulting stress tensors are extracted cell by cell and applied to discrete fracture networks with different orientation statistics. Subsequently, for each fracture within every cell, the normal stress is calculated and used to determine stress-dependent hydraulic apertures through an exponential closure law. This is followed by the computation of fracture transmissivities using the parallel-plate cubic law. Finally, directional hydraulic conductivities are then obtained by numerical flow simulations in the principal directions (Kx, Ky, Kz).The results show strong spatial heterogeneity and directional dependence of hydraulic conductivity across the catchment. At first order, the anisotropy of Kx, Ky and Kz is controlled by fracture network geometry, with low conductivities occurring in directions poorly aligned with dominant fracture orientations (e.g., low Kx for predominantly N–S–oriented vertical fractures or low Kz for horizontal fracture sets). Together with this geometric control, the relative orientation between fractures and the local stress tensor exerts a strong mechanical influence, fractures oriented perpendicular to the principal compressive stress experience increased normal stress, reduced aperture, and consequently lower hydraulic conductivity. In addition, spatial variations in topography introduce local perturbations in the stress tensor that produce zones of relatively higher hydraulic conductivity, where fractures remain less compressed. These combined effects lead to pronounced hydraulic anisotropy and preferential pathways at the catchment scale, highlighting the importance of explicitly accounting for spatially variable stress fields when modeling flow in fractured catchments.

In situ measurements of fracture deformation during fluid injection are rare, yet essential for understanding the mechanical response of fractured rock. In this study, we evaluate the reliability of two methods by comparing their slip vector estimates: high-resolution borehole acoustic televiewer images captured before and after injection tests, and displacement data from a three-component borehole deformation probe recorded during the injections. Acoustic televiewer images capture only final in-plane displacement, whereas three-component borehole deformation measurements provide full 3D, transient fracture movement. Four injection tests in a fractured granitic rock mass along an inclined borehole at the Bedretto Underground Laboratory (Switzerland), beneath more than 1100 m of overburden, were analyzed. The two methods yielded consistent kinematics and comparable slip magnitudes, typically in the range of 0.2 - 0.6 mm. Angular differences between estimated slip directions ranged from 0.2 degrees to 28 degrees. For two test intervals with pre-existing fractures optimally oriented for slip, the measured slip vectors agreed with forward-modeled slip vectors based on the regional in situ stress field, supporting the Wallace-Bott hypothesis. Measurable shear displacement was identified in two test intervals containing fractures oriented nearly perpendicular to the least principal stress, including a newly formed hydraulic fracture. This demonstrates that even fractures unfavorably oriented for slip can undergo shear due to local stress perturbations around an inclined borehole.

The occurrence of springs and their connectivity within stream networks is typically associated with three key controlling factors: climate, topography and the distribution of hydraulic properties. In crystalline media, this distribution is often related to lithology and the presence of fractures. In addition, tectonic and topographic stresses can modify properties through compressive and extensional forces acting on the rock mass and fractures. However, these controls are rarely considered for hillslope scale applications. The aim of this research is to investigate the effects of stress on bedrock hydraulic properties and their implications for groundwater flow and transport at the hillslope scale. A numerical experiment has been designed that combines linear poroelasticity to simulate the distribution of permeability and porosity, together with groundwater flow and transport simulations. Different slope and stress conditions are examined, providing a comprehensive sensitivity analysis framework.Our results show that vertical stress leads to a decrease in permeability and porosity at depth, following an exponential-like trend. Increasing the proportion of lateral stresses relative to the total vertical stresses reduces the mean permeability and porosity and increases the variance in the distribution along the hillslope. For high values of lateral stress, a low permeability domain develops downslope at the valley bottom due to the accumulation of compressive stresses, while the extensive regime at the crest provides higher permeabilities. As expected, groundwater flow simulations revealed that the partitioning of flow paths is strongly influenced by such heterogeneous stress-induced permeability and porosity fields. As stress increases, groundwater flow becomes more channelized in the near subsurface, strongly deviating from the classical Dupuit model. We also found that the distribution of normalized groundwater discharge rates shows higher values in the upper part of the seepage zone than in the lower part. By analyzing the results of particle tracking simulations, we found that mean residence times increase with higher external stress due to a decrease in mean permeability. In addition, the shape of the residence time distribution is strongly modified by the channeling of groundwater flow with increasing lateral stress, with the probability of shorter residence times increasing as stress increases. We discuss the implications of these fundamental results for our understanding of the role of stress in groundwater-dependent systems, with important insights into the recharge, storage and discharge mechanisms that may control the resilience of landscapes to the effects of climate change.

Abstract Groundwater flow at the hillslope scale is controlled by hydraulic properties and topography. Geological media are inherently heterogeneous, with properties depending on both lithology and external factors such as weathering, erosion, and tectonic activity. The effects of stress‐dependent variations in porosity and hydraulic conductivity on groundwater flow and residence times remain unexplored. Using process‐based hillslope models, we demonstrate how external lateral stresses cause significant variations in stress and strain, particularly on slopes steeper than 30%. These stresses reorganize porosity and hydraulic conductivity, with compressive regimes reducing these properties at valley bottoms and tensional regimes elevating values at crests. This redistribution of porosity and hydraulic conductivity leads to a channeling of groundwater flow near the surface. Channeling increases discharge rates in the upper seepage zone and the proportion of rapid flows. Our results highlight geomechanical‐hydraulic property relationships as critical controls on critical zone processes in mountain settings.

We investigate the role of co-seismic devolatilization in driving aftershock sequences, focusing on three recent sequences in Greece: Thiva, Damasi, and Florina. We use a numerical model incorporating non-linear diffusion with a source term that accounts for fluid generation via thermal decomposition of carbonates. Our model successfully reproduces the spatio-temporal characteristics of the observed aftershock sequences, with especially strong correlations between observed and modeled aftershock rates. The calibration of the model, which introduces key parameters for fluid generation and permeability recovery, provides new insights into the mechanisms of aftershock generation. It reveals a consistent magnitude threshold above which a significant amount of fluid is produced. The fluid pressure generated remains within the same order of magnitude regardless of the earthquake size. Our results suggest that co-seismic fluid generation, either through decarbonation or dehydration reactions, is a significant driver of aftershock activity in these regions, and likely globally. These findings highlight the importance of incorporating fluid-related, and fluid-generating, processes into models of seismicity, especially in tectonic environments where devolatilization is likely to occur.

ABSTRACT: From 2022 to 2023, a series of hydraulic stimulation experiments were conducted at the Bedretto Underground Laboratory in Switzerland (BedrettoLab) to study the creation of an engineered geothermal reservoir in crystalline rock. A 400 m long stimulation borehole was divided into 14 intervals using a multi-packer system. During the volume characterization, an azimuthal rotation of borehole breakouts was observed, indicating a stress perturbation caused by a major fault zone intersecting the experimental volume. In this study, we present the seismo-hydromechanical observations from seven hydraulic stimulations in four selected intervals. This is the first systematic comparison of repeated stimulations in different intervals at the BedrettoLab. Two of these intervals are neighboring the fault zone, while the other two are located further away. We utilize data from a dense multiparameter monitoring network to show that the stress perturbation affects the reactivation of similarly oriented pre-existing fractures. In our example, the stress perturbation caused by the fault zone prevents the reactivation of structures in the nearby interval, while similar structures are clearly reactivated in the three other intervals. Reactivation pressures from pressure-flow rate plots agree with the pressures encountered at the onset of seismicity. Given that the prevailing fractures are well oriented for shear reactivation within the far field stress field, hydraulic shearing is considered the probable reactivation mechanism. 1 INTRODUCTION Interest in engineered geothermal systems (EGS) as a low emission, renewable energy source (Lu, 2018; Aghahosseini and Breyer, 2020) has grown since the late 2000s. In Central Europe, EGS reservoirs with sufficiently high temperatures are located at depths of several kilometers, where the permeability of the crystalline basement rocks is insufficient for geothermal energy extraction. Permeability enhancement is achieved through hydraulic stimulation, which can involve hydraulic shearing of natural fractures or shear zones, hydraulic fracturing of intact rock, or a combination of both (McClure and Horne, 2014). This permeability enhancement goes hand in hand with induced seismicity, which can reach damaging levels if large fault zones are reactivated (e.g. Deichmann and Giardini, 2009; Evans et al., 2012; Ellsworth et al., 2019). To address this challenge, several scaled-down in situ hydraulic stimulation experiments have been conducted at decameter scale in underground research laboratories in representative crystalline rock types (e.g. Amann et al., 2018; Zimmermann et al., 2019; Schoenball et al., 2020; Fu et al., 2021).

Seismicity following heavy rainfall events is often considered to be triggered by the pore pressure diffusion front migrating from the surface towards the focal depth, assuming a homogeneous crust. Although this assumption can be justified in some cases depending on local geology (e.g., a homogeneously fractured basement), it is too simplistic for known karst areas. Indeed, the hydraulic behaviour of karst during a flood event is not dominated by pressure diffusion but by highly transmissive karst conduits. This generates rapid and large variations in hydraulic head, and if karstification deepens, this leads to large changes in pore pressure at seismogenic depth. We consider the conditions and data from three different case studies with karstic features to evaluate possible seismicity-triggering mechanisms. We identify four potential mechanisms to explain the influence of rainfall on fault stability: crustal loading and its associated poroelastic deformation, pore pressure diffusion, and direct hydraulic connection. We quantify the effect of these mechanisms for parametrisation corresponding to our case studies by considering the specific features of karst and using simplified analytical solutions. Results show that the pore pressure increase resulting from the crustal loading and poroelastic deformation is much smaller than the pore pressure increase resulting from a direct hydraulic connection and its deepening diffusion over small distances. Moreover, the timing between the intense precipitations and the beginning of seismicity may indicate the mechanism behind the triggering process: no time lag to a time lag of a few hours supposes a direct hydraulic connection, whereas a time lag of a few hours to a few days suggests a coupled solution of a rapid pressure increase resulting from direct hydraulic connection followed by the diffusion process prograding towards the focal depth. Our results highlight the importance of considering the intrinsic properties of karst and its spatial distribution, especially its depth, when studying rain-triggered seismicity in a karstic environment.

Natural groundwater level fluctuation in karstic networks resulting from significant recharge (precipitation and/or seasonal snowmelt) can potentially induce seismicity. Triggering is often considered to be the result of pore pressure diffusion front migrating from the surface to focal depth, assuming a homogeneous crust. Although this assumption can be acceptable in some cases (e.g. homogeneously fractured basement) it is hardly justified in known karstic area. Considering the specific features of karst and data of three case studies, we elaborate a conceptual model of rain-triggered seismicity in karstic regions by identifying potential triggering mechanisms and, using simplified analytical solutions, quantifying their impact on fault stability. Results of this analysis indicate that a direct hydrogeological connection between karstic conduits and the hypocenter can lead to pore pressure variation of the order of MPa, potentially initiating a rupture. To test the conceptual model, field investigations are carried out in the Jura Mountains, a well-known karstic area with low to moderate seismicity. Data acquisition consists in monitoring the natural microseismicity and the flowrate at karstic springs, used as a direct proxy for groundwater level fluctuations.Combining both data sets allows to identify events that are potentially rain-triggered and to acquire a quantitative knowledge on what pressure change, inferred from the hydraulic head increase, is affecting the fault’s stability, a valuable information when planning underground projects.

Deep geothermal boreholes, often drilled to the crystalline basement, suffer from borehole breakouts that compromise borehole stability and/or lead to low drilling performance. These issues increase the cost of deep geothermal projects and lead to irregular cross-sectional geometries that may entangle well completion (e.g., packer isolation for zonal stimulation, cementing, etc.). Thus, the proper knowledge of rock strength, state of stress and their interactions at the closest vicinity of the borehole is key to the success of deep geothermal drilling. Typically, the magnitudes of the vertical and minimum horizontal principal stresses, Sv and Shmin, respectively, can be estimated while SHmax is difficult to constrain. This paper presents a systematic methodology to jointly evaluate the heterogeneous distributions of the stress tensor principal components and orientations, and the rock strength properties (e.g. cohesion, friction). Model parameters are estimated from measurements available during or shortly after drilling, i.e., breakout width, breakout extent/depth of penetration, breakout orientation and drilling induced tensile fractures. Additionally, measurements of estimated parameters or transformations of them can be considered in the calibration in a generic manner (e.g., Shmin interpreted from XLOT). For illustration purposes, the methodology is applied to the extensive borehole data set along the crystalline section of the borehole BS-1, in Basel (Switzerland). The methodology allows us (1) to derive plausible sets of stress and strength parameters reproducing the complex distribution of breakouts along BS-1, and (2) to unveil the paradox of having no borehole breakouts at sections with high density of natural fractures.

ABSTRACT: Overcoring is a common technique for measuring stresses in mining projects. Knowledge of the in-situ stress state is essential to ensure the stability of underground infrastructures as well as to assess the induced microseismic risk associated with deep mining operations. There are different types of overcoring probes. Some are bonded in the pilot hole with an epoxy resin and allow for 3D stress measurement (e.g. CSIRO-HI), and others are based solely on a mechanical coupling of the probe, but are limited to biaxial stress measurement (e.g. USBM). The need to glue the probe to obtain a 3D measurement limits the applicability of this technique to short boreholes because it is technically difficult to glue probes in deep boreholes. In any case, traditionally the data analysis is done only based on the final deformation obtained after overcoring. In this paper we propose to use the transient deformation response during overcoring to: (1) allow to evaluate the 3D stress field from a single biaxial overcoring measurement, and (2) add a quality control component by reproducing the entire overcoring response. The general principle of our approach is to simulate the transient response of overcoring by numerical elastic simulation. The principle of superposition is used to derive the responses for any set of parameters from a limited number of basic models and thus allows to limit the total number of model runs. Such approach allows for a systematic inversion procedure to be applied for determining the optimal parameter sets that best capture the transient response of the overcoring. This allows the estimation of a 3D stress tensor from biaxial measurements. It also allows to have a quality control on the measurements evaluating the quality of the fit between the model and the data. In this paper, we evaluate the robustness of the proposed approach by performing a systematic sensitivity analysis on the stress estimation from the inversion of transient overcoring data. We demonstrate the advantages of the approach but also its limitations. The preliminary results obtained in this study suggest that the transient overcoring response contains sufficient information for constraining efficiently the 3D stress tensor. The inversion must be performed using multiple starting point, and the mode of the obtained calibrated parameters is in close adequacy with expected values, while some outliers can be present in the calibrated set. Further work is required for confirming these encouraging results by testing of broader range of stress configurations and applying the method to actual field measurements.

The scaling laws describing the spatial arrangement of fractures along six deep boreholes penetrating the crystalline rocks in the Rhine Graben were derived using a correlation analysis. Five of the wells, two to 5 km depth, were located at the Soultz geothermal site and one well to 5 km depth was located at Basel, some 150 km from Soultz. Five datasets were derived from borehole imaging logs, whilst one stemmed from the analysis of 810 m of continuous core at Soultz. The two differed inasmuch as the core dataset included essentially all fractures, whereas the image log dataset had few fractures narrower than 1-3 mm. The results of the analysis for all image datasets showed that the spatial arrangement of fractures followed fractal behavior at all scales from meters to several hundred meters, the largest scale amenable to assessment, and that the fractal dimensions were confined to the narrow range 0.85-0.9. However, the core dataset showed significant deviation from fractal behavior, the best-fit fractal dimension of 0.8 being somewhat lower than values obtained from imaging logs in neighboring wells. Eliminating fractures with apertures less than 1 mm from the core dataset to improve comparability led to even lower fractal dimension estimates, indicating the discrepancy was not due to imaging log resolution. Analysis of successive depth sections of the core log suggested the discrepancy was due to the presence of a localized zone between 1750 and 2070 m where the fractal organization is disturbed or takes a lower dimension than elsewhere. Aside from this zone, no systematic variation of fractal dimension with depth was observed in any dataset, implying that a single exponent together with intensity adequately describes the arrangement of fractures along the entire length of the boreholes. The results are relevant to the parameterization of DFN models of deep rock masses.

Abstract High‐pressure fluid injections cause transient pore pressure changes over large distances, which may induce seismicity. The zone of influence for such an injection was studied at high spatial resolutions in six decameter‐scaled fluid injection experiments in crystalline rock. Pore pressure time series revealed two distinct responses based on the lag time and magnitude of pressure change, namely, a near‐ and far‐field response. The near‐field response is due to pressure diffusion. In the far‐field, the fast response time and decay of pressure changes are produced by effective stress changes in the anisotropic stress field. Our experiments confirm that fracture fluid pressure perturbations around the injection point are not limited to the near field and can extend beyond the pressurized zone.

SUMMARY Fluid injections can trigger seismicity even on faults that are not optimally oriented for reactivation, suggesting either sufficiently large fluid pressure or local stress perturbations. Understanding how stress field may be perturbed during fluid injections is crucial in assessing the risk of induced seismicity and the efficiency of deep fluid stimulation projects. Here, we focus on a series of in situ decametric experiments of fluid-induced seismicity, performed at 280 m depth in an underground gallery, while synchronously monitoring the fluid pressure and the activated fractures movements. During the injections, seismicity occurred on existing natural fractures and bedding planes that are misoriented to slip relative to the background stress state, which was determined from the joint inversion of downhole fluid pressure and mechanical displacements measured at the injection. We then compare this background stress with the one estimated from the inversion of earthquake focal mechanisms. We find significant differences in the orientation of the stress tensor components, thus highlighting local perturbations. After discussing the influence of the gallery, the pore pressure variation and the geology, we show that the significant stress perturbations induced by the aseismic deformation (which represents more than 96 per cent of the total deformation) trigger the seismic reactivation of fractures with different orientations.

Engineering an Enhanced Geothermal System (EGS) requires a proper understanding of the fracture network properties from small to large scales in order to create a reliable geological model for reservoir simulations. As deterministic identification of all fractures in a reservoir is practically impossible, stochastic approaches known as Discrete Fracture Networks (DFN) are used. This consists of parametrizing a statistical realization of fracture networks constrained by direct observations from borehole images and/or outcrop data, if available. DFN models can be used to study the thermo-hydro-mechanical (THM) properties of fractured rocks and to simulate the processes associated within: I) fluid circulation, II) flow and heat production as well as III) seismic response to hydraulic stimulations. Fractal DFNs are based on multiscale fracture network characteristics and are constrained by the scaling properties of fracture network attributes such as length (or size) and spatial distribution. The dual power-law model is a mathematical representation of fractures that parametrize fractal DFNs with two scaling exponents: 1) scaling of spatial distribution using two-point correlation dimension of fracture centers in three dimensions and 2) power-law exponent of fracture length distribution. Direct measurements of fracture length exponents from borehole images or cores are an unresolved challenge and the resolution of geophysical investigations is not sufficient to image the natural fracture networks. In contrast, the spatial distribution of fractures may be precisely characterized using borehole image logs and cores. Currently, the depth-dependence of spatial clustering of fracture patterns in the earth’s crust is not fully understood, although it may be required to anticipate deep reservoir conditions from shallower datasets. Here, we study such a depth dependency by using the two-point correlation dimension of fractures along the boreholes as a reliable estimate of the fractal dimension. We investigate the data stemming from two deep boreholes, GPK3 and GPK4, drilled into the crystalline basement rocks at the Soultz-sous-Forêts geothermal site. Recent analyses unraveled no systematic variation of fractal dimension with depth in any of the boreholes at the one standard deviation level of uncertainty. This conclusion may support the hypothesis of generating fracture network models with only a single correlation dimension using the stereological relationships in reservoirs up to 5 km depth in crystalline basements.

This roadmap is the product of a grassroots effort by the Swiss Geosciences community. It is the first of its kind, outlining an integrated approach to research facilities for the Swiss Geosciences. It spans the planning period 2025-2028. Swiss Geoscience is by its nature leading or highly in-volved in research on many of the major national and global challenges facing society such as climate change and meteorological extreme events, environmental pol-lution, mass movements (land- and rock-slides), earth-quakes and seismic hazards, global volcanic hazards, and energy and other natural resources. It is essential to under- stand the fundamentals of the whole Earth system to pro-vide scientific guidelines to politicians, stakeholders and society for these pressing issues. Here, we strive to gain efficiency and synergies through an integrative approach to the Earth sciences. The research activities of indivi- dual branches in geosciences were merged under the roof of the 'Integrated Swiss Geosciences'. The goal is to facilitate multidisciplinary synergies and to bundle efforts for large research infrastructural (RI) requirements, which will re-sult in better use of resources by merging sectorial acti- vities under four pillars. These pillars represent the four key RIs to be developed in a synergistic way to improve our understanding of whole-system processes and me- chanisms governing the geospheres and the interactions among their components. At the same time, the roadmap provides for the required transition to an infrastructure adhering to FAIR (findable, accessible, interoperable, and reusable) data principles by 2028.The geosciences as a whole do not primarily profit from a single large-scale research infrastructure investment, but they see their highest scientific potential for ground-break-ing new findings in joining forces in establishing state-of-the-art RI by bringing together diverse expertise for the benefit of the entire geosciences community. Hence, the recommendation of the geoscientific community to policy makers is to establish an integrative RI to support the ne- cessary breadth of geosciences in their endeavor to ad-dress the Earth system across the breadth of both temporal and spatial scales. It is also imperative to include suffi-cient and adequately qualified personnel in all large RIs. This is best achieved by fostering centers of excellence in atmospheric, environmental, surface processes, and deep Earth projects, under the roof of the 'Integrated Swiss Geosciences'. This will provide support to Swiss geo-sciences to maintain their long standing and internatio- nally well-recognized tradition of observation, monitor-ing, modelling and understanding of geosciences process-es in mountainous environments such as the Alps and beyond.

The in situ state of rock mass stresses is a key design parameter, e.g., for deep engineered geothermal systems. However, knowledge of the stress state at great depths is sparse mostly because of the lack of possible in situ tests in deep boreholes. Among different options, core-based in situ stress estimation may provide valuable stress information though core-based techniques have not yet become a standard. In this study we focus on the Diametrical Core Deformation Analysis (DCDA) technique using monzogranitic to monzonitic rock drill cores from 4.9 km depth of the Basel-1 borehole in Switzerland. With DCDA the maximum and minimum horizontal stress ( S Hmax and S hmin ) directions, and the horizontal differential stress magnitudes (∆ S ) can be estimated from rock cores extracted from vertical boreholes. Our study has three goals: first, to assess photogrammetric core scanning to conduct DCDA; second, to compare DCDA results with borehole breakout and stress-induced core discing fracture (CDF) data sets; and third, to investigate the impact of rock elastic anisotropy on ∆ S . Our study reveals that photogrammetric scanning can be used to extract reliable core diametrical data and CDF traces. Locally aligned core pieces showed similar S Hmax orientations, conform to borehole breakout results. However, the variability of core diametrical differences was large for the Basel-1 core pieces, which leads to a large spread of ∆ S. Finally, we demonstrate that core elastic anisotropy must be considered, requiring robust estimates of rock elastic moduli, to receive valuable stress information from DCDA analyses.

Understanding the complex seismic, thermal, hydraulic and mechanical processes active during the hydraulic stimulation or continuous operation of Enhanced Geothermal Systems (EGS) requires an accurate description of the pre-existing fractures and faults. However, the three-dimensional characterization of the fracture network is challenging, as direct observation of the discontinuity network at great depth is limited. Fracture image logs and continuous core, which provide line samples through the fracture network, are most valuable in this regard as they provide the most precise option to place constraints on network attributes in stochastic realizations of the fractured rock mass. Among various geometrical attributes, the spatial clustering of fractures plays a critical role on the rock mass properties. Here, we analyzed the spatial distribution of fractures derived from image log runs in six deep boreholes in crystalline basement rock. In one well, the fracture distribution from continuous core was also available. The wells were drilled to depths between 2-5 km, and were all located in the same tectonic setting of the Upper Rhine Graben, which is recognized for its high geothermal potential. The normalized correlation integral method was employed to define the scaling relationships of fracture patterns. This methodology is demonstrated to be less affected by the finite size effects, delivering reliable estimates of scaling laws. Detailed analyses of image log datasets revealed fractal scaling with similar fractal dimensions (between 0.85 and 0.96), prevailed over almost two orders of magnitude of scale. The same was true for the fracture distribution derived from the continuous core, although this distribution was found to be more clustered than that derived from image logs in the same well (i.e. the fractal dimension was lower, which may be due to the partial fracture sampling of image logs which have a coarser resolution than continuous core analyses). Analysis of fractures in sub-sections of the core dataset from progressively increasing depths revealed no systematic depth-dependency for the fractal dimension, although a local variation at a scale of hundreds of meters was identified.