Direct shear box tests have revealed that the stiffness and strength of an ice‐filled joint are a function of both normal stress and temperature. Comparison of these data with the results of similar experiments conducted on unfrozen joints indicates that at low temperatures and normal stresses the strength of an ice‐filled joint can be significantly higher than that of an unfrozen joint. However, in the absence of sufficient closure pressure, the strength of an ice‐filled joint can be significantly less than that of an unfrozen joint. This implies that if the stability of a slope is maintained by ice‐filled joints, its factor of safety will reduce with temperature rise. This hypothesis suggests that a jointed rock slope that is stable when there is no ice in the joints and is also stable when ice in the joints is at low temperatures will become unstable as the ice warms. Results from the model tests have confirmed this hypothesis. Copyright © 2001 John Wiley & Sons, Ltd.RÉSUMÉDes tests de cisaillement directs ont révélés que la rigidité et la résistance d'un joint rempli de glace est fonction à la fois de la contrainte normale et de la température (Davies et al., 2000). La comparaison de ces données avec les résultats d'expériences semblables conduites sur des joints non gelés indique qu'à basse température et pour des contraintes normales identiques, la résistance d'un joint rempli de glace peut être plus élevée d'une manière significative que celle d'un joint non gelé. Toutefois, en l'absence d'une pression de fermeture suffisante, la résistance d'un tel joint rempli de glace peut être significativement moindre que celle d'une fissure non gelée. Ceci implique que si la stabilité de la pente est maintenue par des joints remplis de glace, son facteur de sécurité sera réduit avec l'augmentation de la température. Cette hypothèse suggère qu'une pente de roches fissurées qui est stable quand il n'y a pas de glace dans les joints et est aussi stable quand la glace dans les joints est à basse température, deviendra instable quand la glace s'échauffe. Des résultats obtenus par des tests ont confirmé ce résultat. Copyright © 2001 John Wiley & Sons, Ltd.

To assess the safety against failure of rock slopes in cold regions, such as high mountain areas, where stability is potentially maintained by ice in rock discontinuities, the shear strength of ice-filled rock joints was investigated in a series of direct shear-box tests. To permit control and repeatability, the experiments were conducted using simulated rock specimens. These were cast in the laboratory using high-strength concrete. Laboratory measurements showed that at a constant rate of shearing, the interface shear strength between ice and a joint surface of repeatable roughness is a function of both temperature and normal stress.

Time-dependent characterisation of rocks for the entire strain range (i.e. up to and beyond the yield point, where rocks are expected to be fractured) have received considerable attention for improving the long-term stability of deep underground openings. Although extensive experimental studies have been carried out on creep of different types of rocks, very limited studies exist, which investigate intact as well as fractured rock samples taken from the same type of rock. In this paper, the time-dependent behaviour of muddy siltstone was investigated to determine and compare creep properties of intact and fractured rock samples. A series of multistage uniaxial and triaxial creep tests were conducted on the rock samples at room temperature. In addition, multistage triaxial testing was conducted on the rock (intact and fractured) to determine the instantaneous (short-term) stiffness and explore its correlation with creep properties. All stain curves showed an initial instantaneous strain followed by two phases of timedependent strain including transient creep phase (particularly for the first loading stage) and a steady state creep phase. The results indicate that both the instantaneous and creep strain are proportional to the deviatoric stress and confining pressure. This is clearly evident in the fractured rock samples, where larger deviatoric stress resulted in an increased creep strain and strain rate. The relationship between axial strain and time was successfully fitted to Burgers creep model. In comparison with the intact rock, creep parameters (of the Burgers model) for the fractured rock were found to be significantly smaller, corresponding to the larger creep deformation and steady-state creep rate experienced by the fractured rock samples. Despite this difference between the intact and fractured rock samples, the study showed a considerable correlation between the creep parameters of both types of rock samples and their instantaneous elastic modulus (obtained at typical confining pressures). Regression analysis revealed that creep parameters could be reasonably estimated from instantaneous elastic modulus using an exponential function. Furthermore, based on the experimental findings, an improved characterisation of timedependent properties was proposed. We believe this approach provides a good basis for future research to enhance geotechnical modelling of long-term stability of abandoned mines as well as for the application of underground disposal of radioactive waste and oil and gas storage.

The use of geophysical characterization of karst systems can provide an economical and non-invasive alternative for extracting information about cavities, sinkholes, pathways for water infiltration as well as the degree of karstification of underlying carbonate rocks. In the present study, three geophysical techniques, namely, Ground Penetrating Radar (GPR), Electrical Resistivity Tomography (ERT) and Very Low Frequency Electromagnetic (VLFEM) methods were applied at three different locations in relation to fluvial karst, which is listed as an environmentally sensitive area in Rio Vermelho, Mambaí, Goiás, Brazil. In the data acquisition phase, the GPR, direct-current (DC) resistivity and VLFEM profiles were obtained at the three locations in the area. Data were analyzed using commonly adopted processing workflows. The GPR results showed a well-defined lithology of the site based on the amplitude of the signal and radar typologies. On the other hand, the inverted resistivity cross-sections showed a three-layered stratigraphy, pathways of water infiltration and the weathered structures in carbonate (Bambui group). The interpretation of VLFEM as contours of current density resulted from Fraser and Karous–Hjelt filters, indicated the presence of conductive structures (high apparent current density) that might be linked to the weathered carbonate and other conductive and resistive anomalies associated with the water-filled and dry cavities (cave), respectively. The results encourage the integrated application of geophysical techniques such as the reconnaissance for further detailed characterization of the karst areas.

Geophysical methods have a varying degree of potential for detailed characterization of landslides and their dynamics. In this study, the application of four well-established seismic-based geophysical techniques, namely Ambient Noise Interferometry (ANI), Horizontal to Vertical Spectral Ratio (HVSR), Multi-Channel Analysis of Surface Waves (MASW) and Nanoseismic Monitoring (NM), were considered to examine their suitability for landslide characterization and monitoring the effect of seasonal variation on slope mass. Furthermore, other methods such as Ground Penetrating Radar (GPR) and DC Resistivity through Electrical Resistivity Tomography (ERT) were also used for comparison purpose. The advantages and limitations of these multiple techniques were exemplified by a case study conducted on Sobradinho landslide in Brazil. The study revealed that the geophysical characterization of the landslide using traditional techniques (i.e., GPR, ERT and MASW) were successful in (i) the differentiation between landslide debris and other Quaternary deposits, and (ii) the delineation of the landslide sliding surface. However, the innovative seismic based techniques, particularly ambient noise based (HVSR and ANI) and emitted seismic based (NM), were not very effective for the dynamic monitoring of landslide, which might be attributed to the short-time duration of the data acquisition campaigns. The HVSR was also unsuccessful in landslide site characterization i.e., identification of geometry and sliding surface. In particular, there was no clear evidence of the light seasonal variations, which could have been potentially detected from the physical parameters during the (short-time) ambient noise and microseismic acquisition campaigns. Nevertheless, the experienced integration of these geophysical techniques may provide a promising tool for future applications.