1 Introduction

Population growth brings, as a consequence, greater territorial demand and for mineral resources, and increasingly requires the execution of civil works to intervene in the physical environment, such as the construction of slopes in road works, containment works, or in open pit mining operations. Slopes are defined as inclined surfaces constituted by rock and/or soil, of natural origin or man-made, with a view to serving civil or mining works [1, 2]. The rock mass, when excavated for the construction of a slope, undergoes a stress relief process, increasing its exposure to weathering agents that accelerate the reduction of strength and physical integrity of the rock mass and can lead to stability problems, with the occurrence of landslides, failures and falling blocks [3]. Short-term slopes in mining operations are constantly modified and designed with intermediate safety factors, between 1.07 and 1.3. This safety factor defines when a slope is unstable (value less than 1) and it is calculated by the ratio between the soil shear strength and the shear stress of a possible sliding surface in the slope [4]. At the end of the useful life of the mining project, the slopes must acquire a safety factor greater than 1.5 in order to ensure long-term stability and without signs of failure [5, 6]. Although the useful life of mining slopes is short in relation to those present in civil works, the stability of the excavated slopes is one of the main factors for managing the mining activity in order to ensure the balance between operational safety and economic exploration criteria [7].

The most relevant structures that influence slope stability are discontinuities such as joints, faults, fractures and shear zones [8]. Combined to the structuring of the rock mass, information regarding the presence of water in the subsurface, as well as its hydrogeological model, are of paramount importance for the evaluation of its influence on the stability of the slopes from its operational phase until its decommissioning [9]. The flow of water in the subsurface is treated as one of the main triggers for the occurrence of failures in slopes and its existence is conditioned to the presence of fractures and their geometry [10, 11]. The absence of fracture connectivity forms a barrier to infiltration and rapid flow of rainwater that can accumulate and constituting fractured aquifers. The existence of fractured aquifers, combined with the occurrence of intense rainfall events and rock saturation, can result in slope failures [12, 13]. The high rainfall events that are registered in Brazil act as the main triggering agents of failures in soil and rock in Brazil–combined to the geological conditions of the place [14].

Piezometers are the most used monitoring resources for evaluate subsurface water pressures by mining operations. Geological mapping for locating piezometer provides a comprehensive result of the percolation direction of pore-pressures arising from water flow. The determination of installation points for piezometric instruments must be as precise as possible due to the cost of implementation and the inconstancy of the pressures acting on groundwater [15]. A rational alternative for locating piezometers is to carry out previous geophysical studies to recognize the water table level [16,17,18].

Geophysical methods are used to determine the internal structure of a soil or a rock mass from the correlation of physical properties with the results obtained [19]. The application of the geophysical method of GPR—Ground Penetrating Radar uses the principle of transmission of high frequency electromagnetic waves and their response to the existence of shallow subsurface geological structures and features [20, 21]. This method is applied in hydrogeological and geotechnical studies for mapping fractures and faults [22,23,24,25,26,27,28], caves [29,30,31,32,33,34] and mineral research [35,36,37].

The purpose of this article is to present the data obtained from the subsurface survey carried out by the GPR and relate it to the occurrence of fractures, presence of water and, consequently, zones of failures in mining slopes. The Osamu Utsumi Mine was the chosen location for the investigation and has high rates of rainfall events, around 1,700 mm/year. For its effective decommissioning, the geotechnical and chemical stability of the site must be ensured [38].

2 Study area

The Osamu Utsumi Mine—MOU, belonging to Indústrias Nucleares do Brasil, is located on the Alkaline Plateau of Poços de Caldas, in the municipality of Caldas/MG (Fig. 1). Operations for uranium extraction were carried out using the open-pit mining method, with blasting using explosives, during the years 1981 to 1995. The pit is approximately 1.2 km in diameter and from this, approximately 17,200 tons of uranium were removed. Since 1995, the uranium mine has been in the process of decommissioning, but given the conditions of both geotechnical and chemical instability, the site has not achieved effectiveness in this process [39,40,41,42].

Fig. 1
figure 1

Osamu Utsumi Mine Location Map

Full size image

The lithology present in the mine is little variable and consists of ultrabasic rocks, tinguaites, foiaites and nepheline syenites [43]. The basic rocks are found in small dykes that present a degree of alteration both in depth and in surface and in these cases, form lateritic horizons [44]. Foiaites and tinguaites are extrusive and textural varieties of nepheline syenite and present leucocratic characteristics with hydrothermal alteration [45].

The mineralization present in the MOU pit is related to the formation of breccias that occurred 76 million years ago by hydrothermal processes, which caused changes in pressure and temperature. These changes along with thermal solutions and gases caused the generation of breccias with different structural patterns and high permeability, allowing the rise of mineralized hydrothermal solutions [46]. Fluorite, molybdenum and uranium are products of this hydrothermal process, which, due to its intensity, makes it difficult to characterize altered and unaltered zones in the mine [44]. The minerals found in altered zones are potassium feldspar, sericite and kaolinite, characterizing whitish-yellow colors on the exposed faces of the slopes.

3 Methods

3.1 Geophysics

The geophysical method used in this study was the GPR—Ground Penetrating Radar, which locates structures and shallow geological features in the subsurface through the transmission of high-frequency electromagnetic waves. These waves are repeatedly radiated underground by means of a transmitter antenna located on the surface of the equipment [35]. The propagation of the electromagnetic wave depends on the frequency of the transmitted signal and on the electrical properties of the materials, such as electrical conductivity and dielectric permittivity [47].

Dielectric permittivity is related to the ability of a material to store electric charge when subjected to an electric field. Water has a very high dielectric constant compared to soils and rocks. The higher the humidity present in the subsurface, the lower the speed of the electromagnetic signals and the more difficult it is to capture them, that is, the greater the signal attenuation. In general, electromagnetic waves penetrate more easily in sandy and dry rocks, and more difficult in clayey and humid soils [48].

The different electrical properties of the materials present in the subsurface cause part of the waves to be reflected and received through the receiving antenna, also located in the surface equipment. The reflected energy is recorded as a function of travel time (round trip), and is scanned on the equipment display, forming the so-called radargrams. The visualization of the radargrams occurs in real time and, subsequently, there is the possibility of processing the data in specific software, if necessary.

The equipment used in this work was the MALÅ Easy Locator WideRange with a high dynamic range antenna of 450 MHz, used for shallow depth (less than 3 m) and 80 MHz for greater depth, with information displayed in real time on an attached screen (Fig. 2). A total of 900 linear meters were covered, subdivided into 3 levels of benches located on the SE front of the mine pit (Fig. 3). The data acquisition work was carried out in the month of November 2022, whose accumulated precipitation was 209.5 mm [49].

Fig. 2
figure 2

GPR equipment used in subsurface investigation at the starting point

Full size image
Fig. 3
figure 3

Front Benches SE travelled with GPR

Full size image

Through the equipment, a propagation speed value of electromagnetic waves of 130 m/µs was adopted for the entire rock mass, given its small and restricted lithological variety. The data was processed in the cloud in the MALÅ Vision software developed by the manufacturer of the GPR equipment. During the processing, a dielectric constant value of 5 was selected, referring to granites, rock similar to those predominant in the study area. Basic processing filters, such as the AGC (automatic gain control) were adopted to highlight the reflection and refraction patterns of the present fractures by improving the depth resolution [50]. The results obtained consisted of radargrams correlating distance from the initial point of investigation and depth of investigation, both in meters. Later, the radargrams were interpreted and correlated with the descriptions of the slope faces made during the structural analysis of the rock mass.

3.2 Structural analyses

The structural analysis comprised the registration of geological structures and the characterization of the properties of discontinuities. For this, the scanline method was used covering a total length of 100 m in accordance with the geophysical lines of data acquisition by GPR [51]. Data were obtained regarding the discontinuities in relation to their attitude in notation (dipdirection/dip) through the compass [52]. Characteristics related to the degree of alteration and roughness of the walls of these discontinuities were observed based on the Q System for the determination of the friction angle [53]. Structural data were processed, analyzed and interpreted using Dips software (Rocsience). Kinematic analyzes to identify failure modes followed the work of [54, 55].

4 Results and discussion

The analysis of the radargrams defined for each slope analysed interpretative elements that are base on the responses of the rock mass to the transmission of electromagnetic waves. Zones where the waves have been partially attenuated, that is, partial reflection and refraction, were identified as the presence of fractures. Zones with full signal attenuation were identified as presence of water, that is, fractured aquifer. Similar works on the recognition of fractured aquifers based on the attenuation of electromagnetic waves were presented by [28, 56].

The SE Front of the MOU pit presents heterogeneous geotechnical behavior in its different sectors. In this article, the sections with greater representation and better visualization of the strictures will be presented. It should be noted that the resolution of the technique depends on the degree of saturation of the rock mass and in this case, the maximium depth of investigation was 10 m.

The presence of failures was determined in specific points of Sector A and C (Fig. 4A, B) and increase of bulk density in mineralized zone in the Central Sector (Sector B), represented in Fig. 4C. Residual ore zones have greater permeability given the caracteristics of the rock mass and, therefore, become zonas of water accumulation and behave like fractured aquifers. In the Fig. 4D, is remarkable the expressiveness of the failure that covers the entire height of the slope, around 12 m. In Fig. 4E it is visible the greenish coloration of the soil resulted from the weathering and desintegration of the rock and the presence of the clay mineral chlorite originated from the hydrothermal process. Blocks consisting mainly of oxid of magnesium, dark in color, are more resistant to the weathering process and, therefore, break down less easily compared to the materials present around them.

Fig. 4
figure 4

Failures present in the SE Front of the MOU pit

Full size image

The analysis of the radargrams originating in each of the benches on the SE front of the mine, allowed the identification of the contrast originated from the difference in the dielectric constants of the rock mass and water, the probable triggering agent of the present failures. Its possible to observe that all radargrams presented below have a superficial horizontal layer, with less than 1 m deep, with total signal attenuation. This may be associated with the presenc of a discontinuous layer in front of the rest of the rock mass related to relief joints originated from the dismantling of the rock mass when the mine was operational. Combined with this, the rainfall events recorded in the days before the fieldwork.

The radargram of the upper bench (1B) (Fig. 5) identified a strong signal attenuation between 30 and 65 m, due to the existence of water that attenuates the GPR signal and causes it to lose reflection and give rise to signal-absent zones (corresponding to the white areas of the radargram). In this same range of positions on the slope, it was verified the presence of a large extent of rupture originating in the upper portion of the face of the slope. This rupture, nearly 35 m long, is composed of totally friable material with a yellowish color, which indicates the action of weathering in the disintegration of the exposed rock. Because it is sulphide rock, it also has the consequence of releasing sulphides that, when flowing with water, can give rise to acid drainage processes, verified in previous works, such as [40, 42, 57, 58]. With these data, it can be inferred an internal saturation of the rock mass at this point as a possible trigger of the originated rupture.

Fig. 5
figure 5

1B radagram with indications of failures

Full size image

The second bench analyzed, 2B, presents strong signal attenuation along its first 40 m analyzed (Fig. 6). These first 40 m in line are located just below in continuity with the rupture present in 1B and can be characterized as a more fragile zone or, with greater seepage and storage of water in the fractures. The rock mass zones present from 60 m reflect the GPR signal and therefore are presented by darker colors and visually, are characterized by intact rock mass without weathering changes in the primary mineral assembly. Failures are visible both on the face and inside the rock mass, given its continuity represented in the radagram. The presence and intersection of different families of fractures in the rock mass may become preferential paths for water flows and enable the development of wedge failures, according to the stereographic analysis performed (Fig. 7). It should be noted that although the occurrence of wedge failures presents a marginal risk of occurrence, these were verified in the field.

Fig. 6
figure 6

2B radagram with indications of failures, fractures and stables zones

Full size image
Fig. 7
figure 7

Kinematic Analysis for wedge failure of the central sector of the SE front: the blue circle represents the friction cone of the discontinuities. The black dots represent the intersections of discontinuities families

Full size image

The visual contrast observed in the previous radargram is accentuated in the third radargram presented referring to the acquisition of GPR 3B data (Fig. 8). Again, there is a presence of a signal attenuating zone, located between 40 and 60 m, which corresponds to the generalized rupture zone that reaches the entire height of the slope. On this bench, it is possible to observe the signal attenuation also in depth, from 5 m, where the visualization of fractures or hyperboles originated from some anomaly is difficult to identify. From the final 70–100 m of analysis, the reflection of the signal is observed, that is, the response of the dry rock to the electromagnetic wave. In the field, the visual analysis of the face identified a stable rock mass, absent of vegetation, infiltrations and failures. There is still preservation of the marks of the barrels used in the detonation that occurred in the mine for uranium exploration, which indicates that the weathering process is slower in this region of the rock mass, possibly due to the lack of water in the subsurface compared to the other regions analyzed.

Fig. 8
figure 8

3B radagram with indications of failures, fractures and stables zones

Full size image

In the absence of significant failures in the other sectors (Sector A and C) and their greater integrity compared to Sector B, it was possible to identify and correlate visual fractures on the face with linear elements present in the radargram. This indicates the continuity of the fractures towards the interior of the rock mass and, in rock drilling and dismantling processes, whether in mining or civil works, it is extremely important to know this to optimize the fire plan and avoid social and economic damage. It should be noted that the different thicknesses of saprolite existing in the different slopes are decisive to the signal penetration an, consequently, in the identification of fractures.

The 100 linear meters traveled with the GPR in sector A—line 1, indicated through its radargram the presence of a rock mass with punctual zones of signal attenuation, which, in visual analysis of the face, represented places with seepage and weatherd rocks. The presence of intersections of different families of fractures was also verified, identified on the face of the slope and inside the rock mass (Fig. 9). The intersection of fractures can originate water flow paths, as well as cause wedge-type failures, making the rock mass less stable. With the structural data collected for this sector, the kinematic analysis carried out indicated the marginal possibility of occurrence of wedge failures in the rock mass (Fig. 10).

Fig. 9
figure 9

1A radagram with indications of fractures, rockfall and seepage in the face of the slope

Full size image
Fig. 10
figure 10

Kinematic Analysis of Section A: the blue circle represents the friction cone of the discontinuities. The black dots represent the intersections of the discontinuity families

Full size image

The radargram of Sector C showed low attenuation of the signal in depth and indicates high attenuation on surface, probably related to the existence of 1 m of thickness of saprolite allied to the presence of water caused by the rainfall recorded during the days preceding the field work. Close to the surface up to 3 m deep in the final 50 m of the GPR data acquisition, there is a strong signal attenuation, which in the visual analysis of the face of the slope, disaggregated material from the rupture of the saprolite launched was identified. In a more intact face of the rock mass, it was possible to correlate the linear elements present in the first 20 m, with the visually identifiable fractures (Fig. 11).

Fig. 11
figure 11

1C radagram with indications of fractures and rupture

Full size image

5 Conclusions

The use of GPR is a powerful tool to interpretation of the subsurface rock mass, identifying lateral and vertical instability to slope movements. However, depth of investigation is limited compared to other geophysical methods, such as electroresistivity. The three benches of Sector B of the Osamu Utsumi Mine analyzed present similar behavior in relation to the existence of significant failures on their faces. The interpretation of the radargrams correlated these failures to the existence of water stored in fractures inside the rock mass, given the attenuation of the signal in these regions. The correlation between the existence of a fractured aquifer in areas where failures occur emphasizes the need for knowledge and identification of these aquifers in the subsurface for accurately locating water level monitoring instruments. In this sense, the GPR is presented as a tool for lease piezometers in monitoring slopes, whether mining or civil works.

The contrast between the intact face and the broken face was notably observed in the attenuations, or reflections of the emitted electromagnetic waves. The central slopes of the SE front are configured as zones of greater accumulation of water through the fractures and correlated to certain structural patterns, in view of the inclination and lithology present in the slope, they are presented as the most unstable zones of the MOU pit. In addition to the problems with failures, as previously mentioned, there are occurrences of acid drainage in the mine, which generates an environmental aggravation.

In sectors with more intact faces and less expressive failures, which do not compromise the visualization of the rock mass as a whole (Sectors A and C), the GPR was effective in identifying discontinuities that, in addition to appearing on the face, penetrate the rock mass. As said, the identification of fractures inside the rock mass helps in the assertiveness of operations such as planning the detonation of a rock mass in the case of mining and civil works. These fractures may also constitute groundwater flow paths and, in this case, become an important triggering agent for failures.

Finally, the geotechnical and chemical instability observed in the field makes decommissioning work at the mine difficult, especially in this region where rainfall is high. For the effective closing of the MOU, measures related mainly to the superficial drainage of the pit must be adopted, with the adoption of drains in order to avoid the accumulation and percolation of water through the rock mass, which can trigger failures and accentuate the process of generating acid drainage.