1 Introduction

Mine closure occurs as a result of the total extraction of ore reserves within the physical limits of a deposit or the mine area, or because of changes in geological or external economic factors that make reserves unworkable or unprofitable at a given time. Abandoned mines are often associated with marked changes in the surface morphology and hydrogeological regimes, which can lead to contamination and surface instability (Bell et al. 2000; Mhlongo 2023). During the last few decades, environmental problems associated with abandoned mines (Bell et al. 2001; Mkandawire & Dudel 2005; Wilson and Pyatt 2007; Cravotta III 2008; Moreno-Jiménez et al. 2009) have led to regulations in many countries (Kwolek 1999; Ginige 2002; Kroll et al. 2002; Kreft et al. 2007; Rezaie and Anderson 2020; Collyard and Patterson 2021). Currently, closure planning is performed during the earliest stages of mining. The methodology and objectives of mine rehabilitation and restoration have become integral parts of planning applications and licensing procedures (Getty and Morrison-Saunders 2020; Tomlin and Gimber 2023).

Abandoned mining activity has a higher incidence in countries with higher per capita income, where there is also a greater need for land, such as Japan, France, Spain, and Germany, which closed most of their mining operations at the end of the twentieth century. The alterations produced in the environment of an abandoned mine can be diverse, depending on the type of resource exploited, exploitation method, execution (or not) of the abandonment plan, and time elapsed since closure. In recent years, many countries have regulated the closure of mines, compensation for damage, and the rehabilitation of affected areas. For this purpose, they have agencies and programs in charge of all aspects of mine decommissioning (INERIS in France, UK Coal Auhority in the United Kingdom, JOGMEC in Japan, and CANMET in Canada). On many occasions, the reason for the creation of these organizations and programs is the occurrence of geotechnical incidents, such as subsidence, that affect infrastructure and buildings and generate a risk for the population (Trigueros et al. 2021).

Underground mining can directly impact surface relief and the landscape. The greatest disturbance to topography, groundwater, surface water regime, and ecological and economic conditions can be attributed to surface subsidence (Gonzalez-Nicieza et al. 2007; Altun et al. 2010; Marschalko et al. 2012; Galloway 2013; Fathi Salmi et al. 2017). Surface subsidence can be defined as a vertical ground movement and is particularly hazardous for buildings, installations, and services, both at the surface and underground (Galloway 2013; Saeidi et al. 2015; Salmi and Sellers 2022). Surface subsidence may occur gradually, almost imperceptibly, or it may occur quite suddenly. This undesirable phenomenon may affect areas as small as a few square meters or as large as several square kilometers (Carnec and Delacourt 2000; Gonzalez-Nicieza et al. 2007). Understanding the behavior of surface subsidence caused by underground mining is one of the most important issues for the assessment of its environmental impact and appropriate underground mine design (Gonzalez-Nicieza et al. 2007; Alejano et al. 1999; Ambrozic & Turk 2003; Unlu et al. 2013). The risk, type, and magnitude of ground movement depend on the nature of the orebody being mined and the mining method employed (Alejano et al. 1999; Gonzalez-Nicieza et al. 2007). Predicting the magnitude of mining subsidence above massive or near-vertical vein deposits is challenging. In the absence of support, subsidence increases and extends with progressive deepening of the workings. Subsidence over workings in tabular or layered deposits is generally of a lower order and more easily predicted than that over massive or vein deposit workings (Alejano et al. 1999).

Room-and-pillar mining is applied to flat-bedded deposits of limited thickness. This method is used to recover the ore from open stopes. This method aims to leave pillars to support the hanging wall. To recover the maximum amount of ore, the miners aim to leave the smallest possible pillars. The ore contained in pillars is non-recoverable and, therefore, is not considered in the ore reserves of the mine (Hustrulid and Bullock 2001). Pillar design is a key factor for successful exploitation. The verification of the actual pillar dimensions and shape is important because the strength of the pillars largely depends on these factors (Gonzalez-Nicieza et al. 2006; Maritz and Malan 2023). The variability of pillar dimensions is also relevant because if some pillars within the array are significantly smaller than others, these smaller pillars can fail first, transferring stress to neighboring pillars and subsequently triggering pillar collapse (Alejano et al. 2017). In the most critical case, a massive pillar collapse, also known as cascading or dominion failure, may occur when an array of pillars fails in quick succession (Martin and Maybee 2000; Malan and Napier 2011). The collapsing ground and the resulting airblast are hazardous to underground miners and on the surface (Rumbaugh et al. 2023; Al Heib et al. 2023).

Backfilling has been widely used in many mines to assist in managing the stability of mined-related voids (stopes, rooms, or goaf) and to improve the recovery of orebodies (Al Heib et al. 2010; Khayrutdinov et al. 2020; Bokiy et al. 2020; Ile and Malan 2023). The types of filling and their functions and requirements depend on the mining methods, strategies, and sequences (Emad et al. 2015; Shi et al. 2022, 2023a, b; Cui et al. 2023). Several backfilling methods have been used in mining engineering. According to several authors (Grice 1998; Rankine et al. 2007; Shespari 2015), the main types of backfilling are (i) rock backfilling, which uses waste rock generated from mining operations as the main component of the filling material; (ii) hydraulic backfilling, in which high-density slurry is delivered through boreholes and pipelines to the underground workings; and (iii) paste backfilling, which is generated from full-stream tailings.

Shespari (2015) defines rock backfilling as a technology for the transportation of backfill-forming components, such as stone, gravel, soil, or industrial solid waste, using manpower, gravity, or machinery equipment to fill underground mining voids. The gangue obtained from the mine is usually used as a backfilling material, although it has to be crushed, sieved, and mixed because the particle size distribution is one of the key factors affecting the backfilling behavior (Ma et al. 2019). Rock backfilling is an economical method that is feasible and applicable to underground mining. In addition, rock backfilling decreases waste material on the surface, expands usable land on the ground, and decreases environmental pollution by transferring waste rocks to deeper levels out of rain contact, increasing the stability of mined areas, and decreasing surface subsidence (Al Heib et al. 2010; Khayrutdinov et al. 2020; Bokiy et al. 2020; Ile and Malan 2023).

Bodovalle is a siderite mine located in Gallarta (near Bilbao, northern Spain), which is currently in the closure phase after finishing its production in the 1990s. The ore was initially extracted by open-pit mining. Once the pit reached the projected depth, an underground mine was developed by using the room-and-pillar method. The underground mine, which is approximately 2000 m long and 600 m wide with large rooms and intermediate rib pillars, is composed of seven mining sectors. The overburden exceeded 200 m, except in the area closest to the open pit where the overburden was 100 m. Because of several pillar failures during and after mine exploitation, some subsidence processes have reached the surface and have been described elsewhere (Trigueros et al. 2021). The city of Gallarta is located just over some of the rooms of the mine, and a reasonable concern arose when subsidence appeared on the surface. This study aimed to evaluate the global stability of room #1.1 for different backfilling levels because it was one of the rooms with the highest risk of subsidence. As the geometrical and stability criteria defined in closure planning were established more than 20 years ago, new technologies and procedures such as LIDAR technology and numerical methods-based software were employed to evaluate the stability of the rooms and the pillars affected by room #1.1

1.1 Background on room #1.1

Room #1.1 is part of a group of 11 rooms called “Exploited Northwest”. According to the information contained in the geotechnical studies included in Closure Planning (carried out over 20 years ago), the rooms had an average width of 13 m, the intermediate pillars had an average width of 8 m, and the heights ranged from 12 to 33 m.

The geometric model was prepared from the floor and elevation plans of the adjoining rooms, plans of the preparation galleries, and room sections obtained using the laser scanner. Figure 1 shows a plan view of the “Exploited northwest” area. A red square marks the studied area. Previous collapses occurred to the east of the studied zone, between the VS and V2 faults. These collapses occurred because of excessive fracturing of the pillars located east of the fault VS. The entire mine between the VS and V2 faults was backfilled after mining operations were completed because of the recurring instabilities on the area.

Fig. 1
figure 1

Plan view of the “Exploited northwest” mining area and the studied zone

Full size image

As a preliminary study, all intermediate pillars of the mine were studied, and their Factors of Safety were obtained using the tributary area method. Expecting deterioration of the pillars over time because of the flooding of the mine in the event that drainage is interrupted, the Closure Planning established that any intermediate pillar featuring a Factor of Safety (FoS) of less than four required a countermeasure to prevent failure. Because of the risk associated with working in this abandoned mine, backfilling the rooms featuring intermediate pillars below the determined FoS was adopted as the ideal countermeasure.

1.2 Strength of the pillars

The stability of room and pillar mine depends on the stability of the roof of the rooms and on the stability of the pillars. The stability of the roofs was assessed using backfilling. If the roof of a room fails and the instability grows upward, the backfilling stops this growth by inhibiting the movement of the falling rocks. The stability of pillars requires special attention because a single pillar failure may cause a cascading failure.

A Factor of Safety (FOS) of the pillars can be calculated as the ratio of pillar stress to pillar strength (González-Nicieza et al. 2006) (Eq. 1).

$${FOS}_{pillar}=\frac{{\sigma }_{cp}}{{\sigma }_{p}}$$
(1)

where FOSpillar is the calculated Factor of Safety of the pillar, σcp is the pillar strength in MPa, and σp is the pillar stress in MPa.

Pillar stress was obtained using the tributary area method (Brady and Brown 2004). This method was developed by considering a regular array of rooms and pillars (Brady and Brown 2004), but it can be extended to a non-regular distribution of pillars (Arzúa et al. 2014). The load supported by each pillar is that of the column of rock over the pillar and half the width of the surrounding rooms or galleries (Eq. 2).

$${\sigma }_{p}={\gamma }_{r}\cdot \frac{\left(a+c\right)\cdot \left(b+c\right)}{a\cdot b}\cdot d$$
(2)

where γr is the mean specific weight of the rocks over the pillar in MN/m3; a and b are the width and length of the pillar in m, respectively; c is the width of the rooms in m; and d is the depth of the pillar in m.

Several formulae have been developed to estimate pillar strength (Lunder & Pakalnis 1997; Von Kimmelman et al. 1964; Hedley and Grant 1972; Krauland and Soder 1993; Sjoberg 1992; Potvin et al. 1989; González-Nicieza et al. 2006; Panek 1981). Most of these formulae consider the Uniaxial Compressive Strength (UCS) obtained from standard laboratory specimens and a shape factor that modifies the strength of the pillar according to its shape and dimensions.

Geotechnical studies of the Closure Planning used the Panek (1981) method to obtain pillar strength (Eq. 3). It considers the uniaxial compressive strength (UCS) of the rock forming the pillar and the width (W), height (H), and length (L) of the studied pillar (Eq. 3).

$${\sigma }_{cp}=UCS{\left(\frac{1}{W}\right)}^{0.08}{\left(\frac{W}{H}\right)}^{0.41}{\left(\frac{L}{W}\right)}^{0.10}$$
(3)

The dimensions of the studied pillars were obtained from the floor and elevation plans. UCS was obtained in previous studies of the same mine (Trigueros et al. 2021) because it was not possible to enter the mine to recover samples. The test results in these previous studies yielded UCS values of 80 MPa for the intact ore and 60 MPa for the mylonitized (altered) ore. To consider the joints intersecting the pillars, the Geological Strength Index (GSI) of the rock mass was used to fit the generalized Hoek–Brown failure criterion (Hoek and Brown 2019). This criterion yields an average compressive strength of 28.04 MPa. It was proposed that the FoS of the pillars obtained using this method should be greater than four to assess long-term stability according to the closure plan.

Figure 2 shows the Factor of Safety values of the intermediate pillars of the 11 rooms of the “Exploited Northwest” sector according to the tributary area method. As shown in Fig. 2, the pillars between rooms #1.1 and #5.1 obtained values lower than four in most cases.

Fig. 2
figure 2

Factors of Safety of the intermediate pillars in the “Exploited Northwest” room area according to the tributary area method

Full size image

It is to remark that the anticipated mine flooding damage to the pillars mentioned in the Closure Plan is questionable. The mines in the Alsace-Lorraine area (France), on which the hypothesis is based, do not have limestones replaced by siderite with high consistency like those of the Bodovalle mine, but have a mineral locally called “minette,” which is an oolitic ferruginous sandstone. They are composed mainly of limonite with a high chlorite content and are associated with shell fragments (bioclasts) and quartz grains. The elements that compose the mineral are united by a carbonate cement. The average iron content of the mined layers was approximately 30–32%, which is much lower than that in the Bodovalle mine.

After flooding the Bodovalle mine, there is no reason to suspect the generation of an acidic pH that would induce oxidation of the remaining ore because it is a limestone rock with very low sulfide content. The ore of the pillars has a very low sulfide content. The main ore, siderite, may act as a neutralizer under certain conditions; however, under other conditions, it can act as an acid producer (Dold and Fontboté 2001; Dold 2017). Throughout mine operation, from the open pit to the underground mine, no acid mine drainage (AMD) problems have been reported, although the mine is in a rainy area. Siderite is expected to act as an acid neutralizer. Therefore, no AMD generation or degradation of the pillar-forming ore is expected. Thus, the ore in the pillars and possible mine flooding do not affect the backfilling material. In addition, the effect of water flooding in the mine can be beneficial for stability because it will generate an internal pressure in the rooms to support the pillars and roofs (Luo and Yang 2018).

1.3 Evolution of the room #1.1 backfilling and measuring process

Room #1.1 is one of the rooms that was marked as risky because of the value of the FoS obtained by the tributary area method, and it is one of the rooms that has been backfilled. This project aims to study the effect of backfilling on the stability of the room as well as to determine the differences between the projected and actual backfilling processes.

The original projected countermeasure for room #1.1 considered a granular backfill poured through two vertical boreholes drilled from the surface (220–230 m in length). When the backfilling project was developed, room #1.1 was not accessible; therefore, it was estimated that the room had a void space of 12,500 m3, and it was projected to pour 8250 m3 of backfilling, that is, the project intended to backfill 66.5% of the room volume.

The backfilling material is a cohesionless granular material. It is a crushed limestone originating from different excavations in the area. The rocks were crushed using a Kleemann MR130Z mobile impact-crusher machine. The crushed material has a granulometry ranging from 0 to 50 mm (D10 = 6 mm, D60 = 32 mm) with a mean density of 1.75 t/m3 after depositing it, and a friction angle of 45°. The latter also means that the angle of repose of the granular material was 45°. The material was poured into the rooms by means of 650 mm external diameter boreholes with a 500 mm diameter inner pipe of 9 mm thickness.

When starting with the backfilling process of room #1.1 in the year 2022, it became evident that the granular material that was being poured had an angle of repose of 35°, allowing a higher level of backfilling than projected. Based on this difference, it was decided to measure the volume of the room to improve the backfilling process. A mobile laser scanning device based on light detection and ranging (LiDAR) was employed. These devices provide rich geometric information through rapid scanning in the form of a 3D point cloud, which can be used to digitally reconstruct the environment. In recent years, the mining industry has been increasingly exploring the possibility of adopting digital sensing technologies to gain comprehensive information from high-resolution and large-scale imaging (Singh et al. 2023). Laser scanning is employed for many applications in underground mining, such as ventilation, geotechnical monitoring, planning, design and construction, roads and rail haulage, and blasting (Lato and Diederichs 2014; Kukutsch et al. 2015; Campbell and Thurley 2017; Chen et al. 2018; Watson and Marshall 2018; Lee and Choi 2019; Navarro et al. 2019; Singh et al. 2021; Yang et al. 2021; Baek et al. 2022). The use of this technology is primarily justified for safety reasons. Not only are the rooms at risk of collapse, but their current stability conditions are not known; therefore, this is the safest way to obtain the actual geometry of the rooms, with no personnel access to the mine. In addition, it is important to determine the actual volume of rooms to optimize the backfilling operation. The shape and volume of room #1.1 are shown in Fig. 3.

Fig. 3
figure 3

Geometry of room #1.1 obtained by LIDAR from borehole #1

Full size image

The volume of the room #1.1 (9400 m3) was found to be much lower than expected. LIDAR scanning also allowed determination of the exact positions of the backfilling boreholes in room #1.1 (Fig. 4). As shown in Fig. 4, the second borehole was drilled in a smaller zone of the room, and, as a result, the backfilling cones of both boreholes intersected each other. Based on this new information, simulations of the backfilling process were carried out, considering angles of repose of 45° and 35°. The simulation results revealed that backfilling cones would only allow pouring of 3300 m3 or 4200 m3 (35.1% or 44.7% of the total volume), respectively (Table 1).

Fig. 4
figure 4

Location of the two boreholes and dimensions of room #1.1

Full size image
Table 1 Different backfilling scenarios for room #1.1
Full size table

2 Effect of backfilling on stability

The fundamental objective of the backfilling of rooms is to prevent the deterioration of the pillars and failure of the roof. If a roof collapses, the materials above it may also collapse and, eventually could reach the surface, which, in this case, is urbanized. The collapse sequence is interrupted when the developing void space is filled with the broken material. In this scenario, the sponge factor of the broken material plays a very important role. Figure 5 presents a simplified conceptual model of the backfilling to obtain the maximum cavity collapsed ceiling height (Hx) for a given backfilling height (Hr) and sponging factor (FE) considering the original room height (Hc).

Fig. 5
figure 5

Conceptual model of roof caving balance to self-backfill

Full size image

Equation 4 shows the balance of heights in Fig. 5.

$$\left({H}_{c}-{H}_{r}\right)+{H}_{x}=FE\cdot {H}_{x}$$
(4)

By solving Hx of the equation, the roof height that must collapse to prevent further rockfalls can be calculated (Eq. 5).

$${H}_{x}=\frac{{H}_{c}-{H}_{r}}{FE-1}$$
(5)

The Hx values used in the calculation models were obtained by applying this formulation to different backfill heights foreseen in the project. For a room height Hc of 12 m, sponge factor FE of 1.5, and room volume of 9,400 m3, the following results were obtained for the backfill.

  1. i.

    With borehole #1–3200 m3 (34%): the average backfill height Hr is 4.1 m, the sunken ceiling height Hx is 16 m, and the final backfilled height of the model is 28 m.

  2. ii.

    With boreholes #1 and #2–4200 m3 (45%): the average backfill height Hr is 5.4 m, the sunken ceiling height Hx is 13.3 m, and the final backfilled height of the model is 25.3 m.

  3. iii.

    Unbackfilled (Hr = 0 m): the sunken height of the roof Hx was 24 m, and the final backfilled height of the model was 36 m.

These results show that the complete collapse of room #1.1 would not induce a sinkhole in the surface even though no backfilling was considered. To confirm this result and to better understand the effect of backfilling on pillars and rooms stability, a numerical model was developed.

2.1 Numerical model of the rooms and pillars

To better understand the failure mechanism and the effect of backfilling on pillar stability, a finite element code (RS2 by Rocscience) was used. RS2 is a continuous modelling software capable of simulating the effect of stresses on an excavation by evaluating the induced stresses and displacements (Rocscience 2018). For numerical modelling in geomechanics as early as 1988, Starfield and Cundall (1988) pointed out that numerical modelling must not be intended to create a detailed simulation of the world but to better understand the mechanism behind the studied case.

Representative sections obtained from the closure plan and LIDAR scanning were developed. Figure 6 shows the central section of the set of rooms that may be affected by backfilling; in this case, the mean depth of the rooms is 230 m. The model was created with different stages beginning from the non-excavated stage (shown in Fig. 6) up to the stage after roof failure and backfilling of the rooms.

Fig. 6
figure 6

Finite element model of central section of rooms #1.1, #1.2, #2.1, and #2.2

Full size image

The values of the geomechanical parameters for each material used in the model were obtained from previous studies (Trigueros et al. 2021) and are listed in Table 2. Cayuela is a local denomination of a carbonated siltsone (Trigueros et al. 2021).

Table 2 Geomechanical parameters of the materials used in the model
Full size table

It is important to note that when performing finite element modelling, the FoS of the pillars is not equal to that obtained using the tributary area method. Moreover, the FoS varies along the pillar section because the pillars are not uniformly loaded. In fact, the periphery of the pillars is more loaded (higher σ1) and has less confinement (lower σ3), resulting in a lower FoS. This can be seen in Figs. 7, 8 and 9.

Fig. 7
figure 7

Pillar and rooms #1.1 and #1.2. a Maximum rock stresses around the rooms (MPa) and b FoS at the center of the pillar

Full size image
Fig. 8
figure 8

Pillar and rooms #1.2 and #2.1. a Maximum rock stresses around the rooms (MPa) and b FoS at the center of the pillar

Full size image
Fig. 9
figure 9

Pillar and rooms #2.1 and #2.2. a Maximum rock stresses around the rooms (MPa) and b FoS at the center of the pillar

Full size image

Spanish regulations (BOE 1985, 1994) indicate values for the FoS of underground excavations depending on the relationship between the UCS and depth and considering the period of use. These regulations are intended to protect people from accidents caused by instability. On the other hand, the particular regulation regarding the abandonment of mines (BOE 1988) is quite concise, and it only points out the requirement of an abandonment plan to assess the safety of people and goods. The abandonment plan had to be approved by the Spanish mining authority, as was the case in this study. As there are no particular Spanish regulations on the FoS of pillars for an abandoned mine, we the authors, consider that the obtained FoS of the pillars according to the numerical methods are appropriate values for considering rooms safe with moderate pillar heights between 15 and 20 m. Moreover, some simplifications were made; for instance, the width of the pillar between room #1.2 and room #2.1 is actually greater than considered (it reached 20 m, although 17 m was used for the calculations), so the models are conservative and the actual situation should be better. This protects the integrity of rooms #1.1 and #1.2 against possible instabilities in rooms #2.1 and #2.2. Moreover, the plan to allow flooding of the mine will increase the stability of pillars because the water will add confinement pressure to the periphery of the pillars.

2.2 Initial situation, no failure nor backfilling on room #1.1.

The results of the maximum stresses around the studied rooms and the FoS of the corresponding pillars are shown on Figs. 7, 8 and 9. These figures correspond to the initial case, where the room #1.1 has not failed and it is not backfilled. The pillar between room #1.1 and #1.2 is 8 m wide (Fig. 7). The pillar between rooms #1.2 and #2.1 is 17 m wide, which largely reduces maximum stress and enhances stability of the pillar (Fig. 8). Roofs of rooms #2.1 and #2.2 were excavated in cayuelas (Fig. 9), which caused a plasticization and destressing zone extending approximately 2 m over the rooms. This destressing zone reached, to some extent, the left side of the pillar, causing a reduction in the loaded section of the pillar and, consequently, an increase in the stress resulting in the lowest FoS of the analyzed pillars.

2.3 Effect of backfilling on studied rooms and pillars

The effect of backfilling was also studied by means of the Strength Factor (a stress FoS) and the displacements. Three cases were studied (i) the initial situation before backfilling, (ii) with the current 35.1% backfilling, and (iii) with 44.7% of backfilling.

Figure 10 shows the Strength Factor on the roofs of the rooms and on the pillars between the rooms for each of the three scenarios to be compared. Table 3 presents the roofs and pillars Strength Factor for the three scenarios. Both the roofs and pillars Factors of Safety remained almost unchanged by backfilling, except in room #1.1 itself that drastically improved. The remaining rooms and pillars were slightly altered in the same manner for both backfilling percentages.

Fig. 10
figure 10

Strength Factor of the roofs of the rooms and of the pillars between rooms in the cases of a initial case before backfilling, b pouring backfill material only from borehole 1 (current situation), and c backfilling from boreholes #1 and #2

Full size image
Table 3 Factors of Safety of the roofs and pillars of the rooms in the three studied scenarios
Full size table

Figure 11 shows the displacements of the contours of the studied rooms in the three aforementioned scenarios: (a) before backfilling, (b) after pouring the backfill material only from borehole 1, and (c) after pouring the backfill material from boreholes #1 and #2.

Fig. 11
figure 11

Total displacement contours around the studied rooms and values of total displacement on the periphery of the rooms for a the initial case before backfilling, b after pouring backfilling material only from borehole #1 (current situation), and c after pouring backfilling from boreholes #1 and #2

Full size image

Figure 11 shows that the backfilling of room #1.1 causes a greater displacement on the left wall of room #1.2. On the other hand, the right wall of room #1.2 reduces its displacement. In this particular situation, backfilling provokes a movement to the left of all the materials located to the left of the backfilled room.

Table 4 presents the numerical values than can be viewed on Fig. 11. The displacement provoked by this movement may reach some 1.5 mm of extra displacement on the right side of room #1.2 (it is the contiguous room to the backfilled one). Conversely, the displacement helps the left side of room #1.2 by diminishing its total displacement by 0.3 mm. Finally, this movement was almost negligible on the left side of room #2.1 and on both sides of room #2.2. It is also necessary to note that the differences in the displacements were negligible when considering different degrees of backfilling (Table 4). The largest difference occurred in the roof of room #2.2, reaching a value of 0.5 mm between the two scenarios, and this larger difference is caused by the plasticization and destressing zone explained before.

Table 4 Displacements in the periphery of rooms #1.2, #2.1, and #2.2, as shown in Fig. 11, in the cases without backfilling (initial case), with 35.1% backfilling (backfilling only from borehole #1), and with 44.7% backfilling (backfilling from both boreholes)
Full size table

The deformation of the ground surface was also studied using numerical modelling. If the roof of room #1.1 fails but it is not backfilled, the maximum surface subsidence obtained by the model was 1.7 mm right over room #2.1 (approximately the center of the excavated zone). The effect of both backfilling methods reduced this surface subsidence to 1.1 mm of total displacement (mostly vertical displacement). It is important to note that in both backfilling cases, the maximum subsidence location moves to the left of the model; in other words, the backfilling affects not only the amount of subsidence but also the location of the maximum displacement.

3 Discussion

Although previous failures were located to the east of fault VS, and room #1.1 was located to the west of such fault (Fig. 1) (Trigueros et al. 2021), concerns about the stability of this room arose because most of the FoS of the pillars in the studied area obtained by means of the tributary area method were below the FoS considered to be secure for assessing long-term stability. To diminish the danger of a potential collapse, it was proposed to backfill room #1.1. This paper presents the work carried out to assess the effectiveness of backfilling.

Considering that the set of rooms in the studied zone was inaccessible because of the risk of collapse, a laser scanner was introduced through borehole #1 to obtain the geometry and actual volume of room #1.1. Moreover, because the backfilling was already being poured into room #1.1, the laser scanner also served to determine the already backfilled volume and the actual angle of repose of the backfilling material. Using this technique, the filling level of the room from a single borehole (35.1%) was determined to be similar to that projected using two boreholes.

From the more detailed geometry obtained and considering the geomechanical parameters of the different lithologies from previous studies, stability models were created using RS2 finite element-based software. This study focused on three cases, all of which assumed the complete collapse of room #1.1: (i) no backfilling, (ii) current level of backfilling, and (iii) additional backfilling poured from borehole #2. Numerical simulations with RS2 allowed easy evaluation of the mining conditions, making it less costly to conduct systematic analysis than in situ investigations (which sometimes is not even possible to carry out, as is this case) or physical modelling (Yu et al. 2018).

The FoS of the pillars and roofs of the rooms were lower than those established in the geomechanical study of the Closure Plan approved in 2000. Even so, the values obtained were sufficient according to Spanish regulations for active underground work. For instance, a FoS of 1.56 was obtained for the pillar between rooms #1.1 and #1.2 without backfilling, increasing to 2.97 when backfilling from borehole #1 was considered.

The study also revealed that there are differences on the height of the room when considering different amounts of backfilling (from 28.3 m for 34% of backfilling to 25.3 m for 45.7% of backfilling) but this difference does not significantly affect the surface subsidence or the FoS of the pillars.

In the original geomechanical report of the 2000 Closure Plan, a FoS greater than four was proposed to assess long-term stability. This value was based on an analogy between the Bodovalle siderite ore and the ore from the Alsace mines in Lorraine, France, where the ore from the Alsace mines was ferruginous sandstones with limonite oolites, while in the case of the Bodovalle mine, the ore was siderite into limestone with very low sulfide content. This difference makes Bodovalle geomaterials less prone to deterioration. Therefore, the proposed analogy is not valid, and consequently, a lower FoS can be considered acceptable for assessing long-term stability. On the other hand, the planned flooding of the mine would be beneficial for stability because the water induces internal pressures in the rooms, which will help supporting the pillars and roofs of the rooms themselves. In addition, because of the very low sulfide content of the ore, acid mine drainage is not expected to affect the stability of the remaining pillars or interact with the backfill material.

All the studied cases reveal surface subsidence ranging from 1.7 mm in the worst case (no backfilling) to 1.1 mm when considering backfilling. It is important to note that besides reducing the amount of surface subsidence, backfilling also affects the location of the maximum surface subsidence, moving it away from the backfilling location.

It has been demonstrated that the backfilling operation of room #1.1 guarantees its long-term stability and eliminates the risk of subsidence. In addition, simulations carried out using numerical methods also assessed the stability of neighboring rooms. The relevance of this work within the context of post-mining and the management of abandoned underground mines on the stability of urban areas (Clarke et al. 2006; Mert 2019; Kretschmann 2020; Measham et al. 2024) should be highlighted, as it is a pioneering study in Spain. The backfilling of abandoned mines by rooms and pillars that exploit metallic resources is not common (Tesarik et al. 2009; Baotang et al. 2017; Aydan et al. 2023), and is more common in coal mining (Zhang et al. 2015; Zhou et al. 2016; Li et al. 2020; Lai et al. 2024) or in active mines to recover ore-forming pillars (Zhang et al. 2017; Zhou et al. 2019; Feng et al. 2022).

The use of the laser scanner through borehole #1 not only has a great advantage in terms of safety issues but has also made it possible to fine-tune the backfilling operation and help in assessing stability, in view of the large differences obtained with the original closure plan.

Finally, this study is a practical application to diminish the risk of subsidence of an underground mine located under a town, which is a great concern and risk for its inhabitants. If adequate countermeasures are not carried out to correct the potential subsidence at Bodovalle Mine, the collapse of the mine could have an impact on the surface, as has already occurred in some parts of the world, such as China (Yang et al. 2017; Jinhai et al. 2019; Wang et al. 2023), France (El Shayeb et al. 2004; Al Heib et al. 2023), USA (Bell et al. 2000; Fernández et al. 2020), Canada (Samsonov et al. 2014; Sepehri et al. 2017), Greece (Loupasakis 2020), Spain (Herrera et al. 2007), Germany (Harnischmacher and Zepp 2014), Poland (Ciszewski and Sobucki 2022), and Japan (Esaki et al. 2008; Ito and Aydan 2020).

4 Conclusions

Considering the risk of subsidence presented by room #1.1 of the Bodovalle Mine and the consequences it could have on the nearby population of Gallarta (Basque Country, Spain), the overall stability of room #1.1 and neighbouring rooms was evaluated for different backfilling levels. It should be noted that the initial Closure Plan did not consider this post-mining operation, but it was required after a few collapses occurred. It is also necessary to mention that this type of countermeasure is uncommon in this type of abandoned mine.

The studied mine was closed, and access to the mine was prohibited because of the risk of collapse. LiDAR technology allowed to obtain the actual geometry of the studied room, as well as the level of filling, without having to enter the mine. The obtained data from this technology revealed that the actual angle of repose of the backfilling material was lower than expected and that the projected backfilling level using two boreholes had almost been achieved with a single borehole.

The numerical model of the studied set of rooms revealed that backfilling is an adequate countermeasure to diminish surface subsidence and increase the FoS of the pillars, and demonstrated that the backfilling operation assesses the long-term stability of the room, thereby greatly diminishing the risk of subsidence over time. The study also revealed that backfilling affected the location of the maximum surface subsidence, moving this point away from the backfilled zone.

Another interesting conclusion is that there is no significant difference between the analyzed backfilling levels when considering the surface subsidence or FoS of the pillars. This may lead to a procedure for optimization of the backfilling process.