Geophysical-based mineral exploration: A combination of seismic and earth resistivity tomography for expansion of a talc deposit in Austria

Introduction

Raw materials are essential to the functioning and integrity of numerous industrial sectors. Secure access to these materials is crucial to the EU’s competitiveness and economic security. Applied geophysics can play a central role in the exploration of raw materials. Especially in complex geological settings where a simple correlation of well data remains ambiguous, combinations of geophysical methods can provide valuable geological models of deposits.

The Rabenwald talc deposit has been systematically explored in the past through numerous drilling campaigns (wireline core drilling and reverse circulation drilling). The acquired geological information and geochemical data were thereafter used for geological modelling and resource estimation. However, future exploration campaigns will have to be conducted over larger areas and costs of drilling campaigns will increase significantly. In order to reduce drilling costs and improve the planning of the drill holes, the use of geophysical exploration methods is currently under consideration.

In this case seismic and geoelectric methods were selected as the most promising methods to measure physical rock parameters over a wide area and thus draw indirect conclusions about lithological variations, fault zones, and possible mineralisation or talc accumulations. Furthermore, these methods have the potential to identify ore anomalies, which could be taken into consideration during the planning of drill holes and targeted accordingly. In addition, a detailed understanding of the geological structure of the study area and of the deposit could improve geological and geotechnical modelling by combining geophysical methods, outcrop mapping and drilling data.

The study area is situated in Austria, within the Upper Austroalpine Units of the Koralpe-Wölz Nappe System, on the eastern edge of the Alps. Talc mineralisation is associated with a pronounced fault zone that dips at an angle of approximately 5° to the south or south-south-west. The host rocks of the mineralised zone consist, in footwall of granites and granite gneisses, whilst the hanging wall consist of augen gneisses, mica schists, amphibolites and calc-silicate rocks. Talc and chlorite are frequently accompanied by layers of leucophyllite several metres thick, which are mostly associated with orthogneisses. The boundaries between the leucophyllite and the granite gneiss are generally sharp.

In 2026 two intersecting geoelectric profiles (940 m and 580 m) and one 2D seismic reflection profile (420 m) were surveyed (see table 1). The seismic profile is located at the same site as Profile 1 of the geoelectric survey to allow a combined analysis. Although Profile 1 is longer, it was shortened to 500 m length for this illustration to enable comparison with the seismic profile. Figure 1 shows a site plan of the surveys.

The measurements have been performed in winter 2025/2026. The weather conditions where fairly good regarding the contact resistance for the geoelectrics and the coupling on the forest soil was fairly acceptable. For the seismic acquistion the stormy weather conditions affected the signal-to-noise ratio. Besides that, measurement in an active open pit quarry resulted in additional noise due to lorry traffic.

Earth Resitivity Tomography

Earth resistivity tomography (ERT) is a combination of a geoelectrical sounding and mapping survey utilising numerous electrodes as grounding points arranged along a profile (multi-electrode geoelectrics), and a two-dimensional evaluation performed via computer-aided inversion modelling. The electrode array is deployed along an approximately linear profile and connected to the measuring instrument. Each electrode can be specifically addressed and switched to function either as a current-injecting electrode (A or B) or as a potential-measuring electrode (M or N). Controlled by the instrument, a current is injected into the subsurface via two electrodes (A and B), while the resulting voltage drop is measured across at least by two other electrodes (M and N). Consequently, four electrodes are always active during any single measurement. Depending on the subsurface resistivity distribution, the injected current intensity, and the currently active electrodes, a characteristic electrical potential difference develops between electrodes M and N. The instrument determines the apparent electrical resistance from the measured voltage, the injected current strength, and a geometric factor (K) that depends on the array type, spacing of the active electrodes and the topography. Step-by-step expansion of the electrode array separation enables the measurement of current samples at increasingly deeper target depths. By gradually moving the active 4-electrode array along the line, the lateral resistance distribution in the profile direction can be determined. In this project, a dipole-dipole configuration was used. An 8-channel measurement system allowed multiple measurements to be recorded simultaneously, thereby significantly reducing measurement times.

The collected raw datasets are subsequently evaluated. They are inverted into a realistic 2D subsurface model of resistivity distribution using a two-dimensional inversion scheme. Surface topography influences current flow fields and must therefore be explicitly incorporated into the processing and inversion workflow. Through inverse modelling of the dataset, the program generates a two-dimensional starting model representing the initial electrical resistivity distribution. In a second step, a forward modelling run (based on the finite difference method) calculates the resulting theoretical apparent resistivities of this model, which are then compared against the actual field-measured apparent resistivities. Subsequently, an optimised model is calculated, and the synthetic data is matched against the field data again. This iterative optimisation loop (iteration) is repeated until the mean data misfit error converges towards a specified threshold or a predefined maximum number of iterations is reached.

Typical resistivities for the rock types relevant to the study area fall within the following ranges:

• Talc (solid, compact): approx. 300 to 7000 Wm; depending on purity, porosity, and degree of weathering and water conent
• Gneiss: approx. 300 – 1,000,000 Wm; depending on purity, porosity, and degree of weathering and water conent
• Kornstein / quarzite rocks: approx. 9000 to 90,000,000 Wm; depending on purity, porosity, and degree of weathering and water conent

When visualising geoelectrical measurement results, lithological boundaries and layer interfaces are typically not expressed as sharp boundaries but are instead visible as gradual resistivity gradients, displayed as ‘Colour transitions; in the cross-sections.

Moreover, geological interpretation can be non-unique/ambiguous.

Seismic reflection and tomography

In seismic surveys, the propagation velocity of compression waves (P-waves) plays a key role in differentiating lithological units. Typical P-wave velocities for the rock types relevant to the study area fall within the following ranges:

• Talc (solid, compact): approx. 3.5-5.5 km/s; depending on purity, porosity, and degree of weathering; fine-grained aggregates may exhibit lower values
• Gneiss: approx. 3.8-6.5 km/s; highly dependent on mineral composition and textural anisotropy
• Carbonate rocks (limestone, dolomite): approx. 3.5-6.5 km/s; high values in dense, low-porosity varieties
• Kornstein / quarzite rocks: approx. 5.0-6.5 km/s; relatively homogeneous and often high moduli of elasticity
• Graphite-bearing zones or fault zones: approx. 2.0-5.0 km/s; significantly reduced speeds due to loosening, increased porosity, and material inhomogeneity

Wide ranges of values and, in some cases, relatively small differences in velocity between the talc deposits and the surrounding bedrock underscore the need for integrated geophysical exploration.

The acquisition spread comprised a total of 212 active recording channels, ensuring high lateral coverage and sufficient redundancy of the acquired seismic data. Data acquisition was performed, by detonating small explosive charges through this fixed spread at every second receiver station. The sensors used were single geophones with a natural frequency of 10 Hz, which were deployed in direct contact with the ground at the defined receiver stations. These types of geophones are particularly well-suited for near-surface reflection seismic applications as they exhibit sufficient sensitivity in the relevant frequency spectrum. The average distance between the geophones (receiver stations) positions was two metres, which enables high lateral resolution and detailed mapping of seismic discontinuities. The seismic signal was generated by controlled blasting using commercial explosives. The charges used ranged from 30 g to 125 g, adjusted to the respective subsurface conditions and the desired signal energy. The explosive charges were located in prepared boreholes at a depth of approximately one metre. The charges were initiated using electric detonators to ensure precise timing of the blast and, consequently, high-quality seismic signals. The average distance between the individual shot points was four metres, which ensured sufficient seismic fold and enabled reliable seismic data processing particularly with regard to velocity analysis and interpretation. This dense source geometry contributes significantly to improving the signal-to-noise ratio and the lateral resolution of the seismic profiles.

The seismic processing was carried out in several phases. On the one hand a standard time-domain processing was done followed by a depth conversion of the result. On the other hand a prestack depth migration (PSDM) was done in addition. The result is already in the depth domain. First breaks were identified and picked semi-automatically in the raw field data sets to calculate static correction values and to be able to carry out a so-called turning ray tomography (Zhu et al., 1992a,b; Stefani, 1993). First, a one-dimensional initial model was created for this pick dataset. For this initial model, synthetic travel times were calculated for each shot point and each receiver. Residual times were then determined by comparing the observed and synthetic travel times. These residuals were fed into the tomographic inversion. The time domain processing was carried out by the following main steps: refraction statics; spherical divergence correction; zerophase spiking deconvolution; surface wave noise attenuation; air blast attenuation; despiking ; bandpass filter (20-130 Hz); spectral whitening (20-130 Hz); F-K velocity filter; velocity analysis; dynamic corrction (NMO); CDP stacking; and F-X deconvolution.

As this particular case involves complex geological structures, a prestack FD shot depth migration (PSDM) was also carried out, as this can yield very good results in such cases. For this process, the preprocessed shot gathers are further conditioned as appropriate, and then each trace is migrated before stacking. The migrated shot gathers are then stacked to generate the final, depth-migrated section. In this process, the CMPs are redefined at intervals of two. This procedure resulted in 212 traces instead of the previous 423 traces per section. The main steps of the depth processing are: Importing raw data records, including populated trace headers; signal processing; editing the relevant data ranges (muting); prestack FD shot-depth migration where migration velocity derived from tomography; bandpass filtering; and ensemble stacking.

Combinded geological interpretation

The geology posed challenges for both methods. On the one hand, the resistivity contrasts in the local lithologies are very small (Palacky,1987, Loke, 2004). On the other hand, solid-rock seismic surveys are challenging due to low impedance contrasts. Therefore geoelectric and seismic data were interpreted independently of one another, and the results were ultimately compared. If the interpretations yield similar results, the confidence in the results is high. Of course, this combination was only possible on Profile 1, where both methods were measured.

ERT: Profile 1 (see Figure 2) shows a differentiated, high-resistivity area at the beginning, close to the surface, which can be correlated with geological faults. At station meter 360, a distinct low-resistivity area is recognisable, potentially indicating talc. External resistivity measurements of local lithologies have shown that conductivity is strongly influenced by water content. Consequently, talc was also found to correlate with low resistivity values, even though it is barely conductive when dry.

Seismic: The fault interpretation focused primarily on the footwall area, since offset values and discontinuities in the reflection patterns at the upper edge of the footwall are generally most clearly discernible due to the pronounced reflective character of this zone. Based on the fault structures identified there, these were traced upward into the hanging wall areas of the profile, where reflections become progressively weaker. Consequently, the seismic data could not provide any direct evidence of trough formation, since the seismic signal was too weak near the surface.

The interpreted structures suggest the presence of a thrust zone accompanied by several antithetic faults. The observed fault pattern suggests complex tectonic deformation with both compressional and local extensional components, with the antithetic faults interpreted as accompanying structures within the deformed thrust system.

Overlaying the results of both measurements match well in the tectonic aspect (see Figure 4). The internal structure of the high-resistance zone at the beginning of the profile correlates with the faults interpreted in the seismic data. Figure 5 shows a detail of the structural indicators in the ERT.

Since both methods show anomalies at the same location, no measurement artifact is expected. The structural geological interpretation appears conclusive and will be incorporated into the deposit model. The geoelectric indications of talc still need to be verified.

References

• Loke, M., and Lane, J. [2004] Inversion of Data from Electrical Resistivity Imaging Surveys in Water-Covered Areas. Exploration Geophysics, 35, 266-271. Doi: 10.1071/EG04266.
• Palacky, G. [1987] Resistivity Characteristics of Geological Targets. In: Nabighian, M., Ed., Electromagnetic Methods in Applied Geophysics-Theory, Society of Exploration Geophysicists Tulsa, OK, 53-129.
• Stefani, J.P. [1993] Possibilities and Limitations of Turning Ray Tomography: A Synthetics Study. SEG Annual Meeting, Extended Abstracts, 610-612.
• Zhu, X., Sixta, D. and Angstman, B. [1992a] Tomostatics: turning-ray tomography + static corrections. 62nd SEG Annual International Meeting, Expanded Abstracts, 1108-1111.
• Zhu, X., Sixta, D.E. and Angstman, B.G. [1992b] Tomostatics: Turning ray tomography + static corrections. The Leading Edge, 11(12), 15-23.