Introduction

Large scale open cut mining operations have resulted in a significant increase in the volume and size of steep waste dumps. Active iron-ore mining operations currently cover 5% of Goa (or approximately 18,300 ha) (Chaulya et al. 1999). Throughout India, iron-ore is predominantly mined using open cut mining methods. The present iron-ore mining rate of 16–21 million tonne per annum coupled with high ore-overburden ratios of around 1:3 results in a need to remove and dispose of 40–50 million m3 of waste per year. The waste dumps generated by this level of disposal often comprise of not only overburden material, but sometimes processed tailings. The development and management of such facilities can pose major problems in terms of mine operations and later, the mine reclamation.

This paper presents a detailed geotechnical investigation of the stability of mine waste dumps and a review of the relevant geo-environmental issues considering high rainfall effects. By obtaining actual site data, it has been possible to perform a limit-equilibrium slope stability study of a mine waste dump. This study has quantified the effect of various soil properties and an increase in pore pressures due to heavy rainfall on waste dump stability.

Literature review

Waste dump instability: a geotechnical perspective

The instability of large waste dumps can result in both operational and environmental problems. Due to increased environmental awareness and the risks to personnel and equipment, mine operators are now making efforts to address both short-term and long-term concerns.

Robertson (1986) has ascertained that the long-term stability of waste dump slopes can decrease as a result of (i) increases in the groundwater table (due to groundwater accumulation and changes in the permeability of the dump materials resulting from weathering and the in-washing of fines), and (ii) decreases in the dump material strength (due to weathering). Robertson’s general findings were supported by the Indian Bureau of Mines (IBM 1995) wherein it was reported that inadequate drainage increases the phreatic surface ultimately resulting in decreased shear strength of the underlying material. Furthermore the IBM (1995) reported that improper compaction decreases shear strength, increases permeability, and increases excessive consolidation; with differential settlement eventually leading to failure of embankments. Caldwell and Moss (1981) proposed (i) limiting the height of dumps, (ii) flattening slopes, (iii) avoiding weak foundations at the toe of the dump, and (iv) preventing pore pressure build-up through drainage as a means to ensure long-term slope stability.

The idea of rainfall infiltration of a slope has been explored by several researchers. Through three-dimensional numerical modelling the effect of various rainfall patterns on the groundwater response of an unsaturated slope has been explored (Ng et al. 2001). Prolonged rainfall with high rain volume produced the greatest change in pore-water pressure and it was noted that high intensity did not necessarily produce a dramatic increase in pore-water pressure (here surface permeability prevented complete infiltration resulting in significant run-off). The general findings of this research are supported by Gavin and Xue (2008); and Xue and Gavin (2008) who assessed the rainfall infiltration into unsaturated and partially saturated slopes, respectively. Their findings indicated that not only will the rainfall intensity and duration control infiltration, but also the rainfall pattern. Greater infiltration was seen when rainfall was initially of high intensity and later of low intensity (as opposed to events that commenced with low intensity and concluded with high intensity).

Cho and Lee (2001, 2002) considered not only infiltration characteristics, but also the resulting stability of a slope due to rainfall. Their research concentrated on shallow failures induced by infiltrating rainwater which created zones of reduced matric suction and thus reduced shear strength. However, this paper will consider deep-seated failure resulting from long rainfall durations with high intensities (monsoonal rains). Under these circumstances Ng et al. (2001) has shown that even after rainfall has ceased the hydraulic head at depth will continue to rise thereby increasing the probability of deep-seated failures.

Waste dump instability: a geo-environmental perspective

Mine waste containment and dump stabilisation are vital for successful mine site rehabilitation. The recent geochemical research studies conducted on mine waste from Goa by Yellishetty (2004); Kumar et al. (2003); Juwarkar et al. (2003) and IBM (2002) emphasise this need for the mining industry of Goa. Juwarkar et al. (2003) performed chemical analysis on mine waste typical of the area; these findings are presented in Table 1. The high heavy metal content of the waste material is of key importance when considering the environmental impact of a waste dump slope failure. Further mineralogical studies by Kumar et al. (2003) suggest that ore commonly found in the State of Goa comprises of hematite, magnetite, limonite, and goethite. Further petrographical studies conducted by them identified additional ore mineral phases not listed above including; pyrrhotite, platinum bearing minerals (e.g., marcasite and magnemite), and gold bearing arsenopyrite. These sulfide minerals have potential acid-producing characteristics. In general, the pyrite oxidation depends upon: surface area of pyrite, form of pyretic sulfur, oxygen concentration, solution pH, catalytic agents, flushing frequencies, and the presence of Thiobacillus bacteria (Caruccio et al. 1998; Smith and Shumate 1970).

Table 1 Physico-chemical characteristics of mine waste from Goa
Full size table

IBM (2002) reports that agricultural fields, irrigation canals, river beds and creeks are prone to heavy siltation and sediment deposition in Goa. This is due to the heavy in-wash from mine dumps occurring annually after periods of heavy rainfall. It was reported by IBM (2002) and Juwarkar et al. (2003) that mine waste is deprived of plant supporting nutrients, hence posing a great threat to agricultural yields and fish populations.

Yellishetty et al. (2008) has shown the potential of using this waste in civil engineering construction as an alternative to the predominantly granite and laterite aggregates currently utilised in construction materials.

Location, geology and climate of the study area

Samples used in testing were obtained from Sesa Goa Limited’s Codli Mine, Goa. The landmass of Goa is a rising plateau between longitude 14°48′00″ and 15°48′00″ and latitude 73°40′00″ and 74°20′07″ with an area of approximately 3,700 km2 and a coastline extending over 100 km (Fig. 1). Goa produces large quantities of iron and manganese ore, while bauxite and silica sand are also extracted.

Fig. 1
figure 1

Location map and plan of the study area at the time of the study

Full size image

The iron-ore formation of Goa is considered to be the northern most extension of the Shimoga-North Kanara-Goa supracrustal belt of the Dharwad super group area and forms the western part of the castle rock schist belt. The principal iron-ore ranges of Goa are: (1) Bicholim–Pale range, (2) Conquirem–Poikul–Melca range passing through Sacorda, (3) Shigao–Kaley–Tathodi range, (4) Kirlapal–Codli–Costi range, (5) Tudou–Barazan–Uguem range, and (6) Pirla–Sulcorna range. The current study area of Codli is located in Sanguem Taluka, South-Goa district, Goa (Fig. 1).

The study area receives rainfall from southwest monsoons between the months of June and September. The average rainfall varies between 3,000 and 4,000 mm with highest rainfall occurring in July while the lowest rainfall occurs in February.

Materials and methods

Soil sampling and testing

Overburden samples were collected from the six waste dumps marked in Fig. 1. In collecting these samples, the procedures recommended by Sobek et al. (1978) and ASTM D4220 (ASTM 2007a) were followed. Each representative sample weighed 3 kg. Care was taken to retain the soil moisture content by sealing the samples in plastic bags. Later, an array of laboratory experiments was performed on each sample (Table 2). Cylindrical specimens with a diameter of 38 mm and height twice the diameter were used in the triaxial testing.

Table 2 List of experiments conducted in the study and relevant standards followed
Full size table

Ground water monitoring

To evaluate the effect of ground water on geotechnical stability, the water regime of the study area was observed. Monitoring of the groundwater table was carried out at dug-wells 1–5 (DW 1–5, Fig. 1) where water levels were measured prior to and during the monsoon season for a period of 4 years. The selection of dug-well positions was based on the underlying aquifer location, accessibility and distance from mining operations.

Numerical modelling

Factor of safety (FOS) analysis was carried out using the limit equilibrium analysis program SLOPE/W (Geo-Slope International). The comprehensive formulation of SLOPE/W makes it possible to easily analyse both simple and complex slope stability problems using a variety of methods to calculate the FOS. This study utilised the general limit equilibrium (GLE) theory. GLE solves two FOS equations: one satisfying force equilibrium and the other satisfying moment equilibrium (a more detailed description of these can be found in Krahn 2004).

The GLE model incorporated an automatic search routine to calculate the critical slip surface and minimum FOS. It was assumed the dump foundation was horizontal and stable. The waste was modeled as free-draining, but the effect of different phreatic surface levels was also explored to assess the effect of pre-monsoon and monsoon season water table levels. Dump heights of 10, 11, 13, and 15 m were modeled using a slope angle of 33° (equal to the average angle of repose) and material properties corresponding to densities of 1.5 g/cc (i.e., c = 12 kPa, ϕ = 13.82°), 1.92 g/cc (i.e., c = 14 kPa, ϕ = 28.89°), and 2.3 g/cc (i.e., c = 26 kPa, ϕ = 38°).

Results and discussion

Material properties

The soil samples collected from the study area were analysed in the laboratory to obtain their moisture content, specific gravity, bulk-density, maximum dry density and shear strength. These results are reported in Table 3.

Table 3 Physical characteristics of the waste dump material from study area
Full size table

Particle size distributions of the dump material were found for the six samples (Fig. 2). The particle size distribution reported varies greatly from those presented for mines located in other parts of the world (i.e., Davis and Ritchie 1987; Rassam and Williams 1997; Williams 1996). The obvious variability in the particle size distribution between each sample and those presented by other researchers is due mainly to differences in crushing and compaction during extraction and dumping of the waste, and the waste dump material’s initial inherent geological properties. Figure 2 shows that approximately 20% of the material is within the sand–silt–clay range. Often it is believed that the unsaturated clay fraction of mine waste may lead to dump slope instability (Chu et al. 2002) and this may be one of the factors contributing to dump instabilities in Goa at heights that prove stable elsewhere. According to the unified soil classification system (USCS) all the samples collected fall into the SW soil category (well graded sands, gravelly sands, little or no fines) (Table 4). However, the uniformity coefficient varies greatly between samples.

Fig. 2
figure 2

Particle size distribution of six samples taken from the Codli Mine waste dumps

Full size image
Table 4 Classification of the Codli Mine waste dump soil in accordance with USCS
Full size table

Unconsolidated–undrained (U–U) triaxial tests were used to determine the shear strength and cohesion of the waste dump material at the field bulk density (1.5 g/cc). A plot of deviator stress versus axial strain was used to compute values of normal stress (Fig. 3). From these results the Mohr’s circle was constructed and the cohesion (c) and angle of internal friction (ϕ) were measured as 12 kPa and 13.82°, respectively for the material’s field bulk density (1.5 g/cc).

Fig. 3
figure 3

a Stress strain curves obtained through triaxial testing on the Codli Mine waste dump material at field bulk density (density = 1.5 g/cc). b Stress strain curves obtained through triaxial testing on the Codli Mine waste dump material with increased density (density = 1.92 g/cc). c Stress strain curves obtained through triaxial testing on the Codli Mine waste dump material with increased density (density = 2.34 g/cc)

Full size image

In order to further understand the consolidation behavior of the dump material and its effects on stability, subsequent U–U triaxial tests were performed on the waste dump material at densities of 1.92, 2.0, and 2.3 g/cc (increased densities were obtained through increased compaction) (Table 5).

Table 5 Relationship between density, cohesion and friction angle of the Codli Mine waste dump material
Full size table

Groundwater monitoring

The groundwater table in the study area is at 32 m with respect to mean sea level (MSL). The current mining operations have gone below −24 m (MSL). Variations in groundwater depths obtained from observation wells (during pre-monsoon and monsoon seasons of different years) are shown in Table 6. The readings demonstrate how relatively close the water table is to the ground level in the study area. The ground water height bears considerably on any geotechnical works in the area and is of great concern to mine management with high phreatic surfaces leading to mine slope failures in other areas of Goa. Also of great importance and relevance to this study is the dramatic change between pre-monsoon and monsoon levels, which can vary between 1.25 and 9.1 m. Changes as dramatic and large as these will clearly influence dump slope stability. This paper aims to quantify at what point these changes may lead to slope failure.

Table 6 Ground water level during pre-monsoon and monsoon seasons
Full size table

Stability analysis of the waste dump from study area

The stability of the waste dump was modeled using GLE methodology for various dump heights and with different ground water conditions. Increases in the ground water level were used to approximate changes that are induced during the monsoon season. Through well monitoring and consideration of the work of Ng et al. (2001) a rise in the water table is inevitable during the monsoon season given the rainfall intensity and duration. Figure 4a shows a typical failure surface for a waste dump from the study area without pore-water pressure, while Fig. 4b shows the effect of a rise in the water table. The scenarios modeled in this paper only consider the dumps under steady state conditions. The effect of a transient rise in the water table are outside the scope of this paper, although it is known that infiltration of rainwater through an unsaturated zone results in the formation of a wetted zone near the slope surface which may lead to shallow failures during periods of rainfall (Cho and Lee 2002; Fourie et al. 1999; Rahardjo et al. 1994).

Fig. 4
figure 4

Limit equilibrium failure envelope of a waste dump generated in SLOPE/W: a without PWP and b with PWP

Full size image

A complete set of results obtained through a parametric study using SLOPE/W is shown in Fig. 5 which outlines the FOS of different dump heights and PWP conditions. Considering the field bulk density (1.5 g/cc) under non-monsoon circumstances (without pore pressure), and with a dump height of 13 m the FOS was unity. However, once the dump height exceeds 13 m, the FOS falls below unity (Fig. 5). Furthermore, even if the dump height is reduced to as low as 10 m, it still becomes unsafe if the PWP exceeds 2.9 m, a common event in the area during monsoon season (Table 6). From an operational perspective, given the heavy rainfall experienced in the region and the high sensitivity of the FOS to changes in the phreatic level, active monitoring of PWP in dumps is clearly necessary.

Fig. 5
figure 5

Results of parametric study completed using SLOPE/W showing the relationship between the dump height and FOS under various PWP conditions [1.5 g/cc bulk density (c = 12 kPa, ϕ = 13.8°); 1.9 g/cc bulk density (c = 14, ϕ = 28.9°); 2.3 g/cc bulk density (c = 16, ϕ = 38°)]

Full size image

The experimental results presented in Table 5 show that the friction angle and cohesive strength of the soil increase with an increase in density as a result of compaction. Further numerical modeling revealed that when the dump was modeled using the cohesion and friction angle associated with the compacted material, the FOS increased dramatically (Fig. 5). In terms of dump management and development, this increased compaction could be generated through the movement of heavy earth moving machinery over the dump. This increased compaction will also aid in limiting environmental impacts of a waste dump, with an increase in cohesion aiding the prevention of soil erosion and dump collapse (Wu 1995).

Based on the results of numerical modeling conducted using SLOPE/W and discussed above the following empirical relationship was developed to predict the FOS of a waste dump (Eq. 1):

$$ {\text{FOS}}\,\, = \,\,0.027c\, + \,0.04\phi \, - \,0.033{\text{DH}} - \,0.083{\text{PWP}}\, + \,0.65 $$
(1)

where FOS is the factor of safety, c is the cohesion (kPa), ϕ is the angle of internal friction (in degrees), DH is the dump height (m), and PWP is the pore-water pressure (m) Multiple linear regression analysis of a dataset comprising four variables which included 50 data points each were used to establish this empirical relation. The model shows a good fit to the numerical modeling data (r 2 = 0.99). Further studies are required in order to test its applicability to waste dumps of Goa more generally.

An unstable dump slope has the potential to create a number of environmental issues. If the slope is structurally unstable (as discussed in this paper), or its surface is prone to soil erosion, transportation of mine waste to off-site areas may occur; this is a situation operators are keen to avoid. The latter issue is generally addressed by surface stabilisation measures.

The importance of maintaining stable slopes becomes clear when considering the work of Juwarkar et al. (2003), Kumar et al. (2003) and Yellishetty et al. (2009) into the heavy metal and acid-producing mineral content of mine waste found in Goa (Table 1). If this waste, devoid of plant-supporting nutrients and carrying high concentrations of heavy metals, is washed into the local area the sequential social and environmental effect would be devastating.

Conclusion

The effect of heavy monsoonal rainfall on waste dump stability has been assessed. It has been shown through numerical modeling that the influence of monsoon rainfalls on the factor of safety of a dump slope is very dramatic (Fig. 5). When free flowing overburden material is dumped, it was observed that a slope angle of 30–35° was formed. An empirical model has been developed to help predict the occurrence of deep-seated waste dump slope failures caused by monsoonal rains and the resulting raised water table. This paper shows that an increase in compaction of the material during the construction of the dump will lead to dramatic increases in the stability of the dump slope. The stability of a waste dump is crucial not only to prevent personal injury or death, asset loss or damage, but also detrimental downstream environmental effects.

The possible environmental impacts of a dump collapse have been summarised. With some dumps containing high heavy metal concentrations and acid-producing minerals a failure to contain them within a stable dump can poison downstream water supplies and dramatically effect crop yields. Although further research is required to examine the stability of mine waste dumps in monsoon affected areas, it is hoped that this research and the empirical relation developed will serve as a rapid indicator of potential slope instabilities and demonstrate the effectiveness of increased compaction of waste material during dump construction.