Archived: Interstate Technical Group on Abandoned Underground Mines
Fourth Biennial Abandoned Underground Mine Workshop
Underground Mining and Its Surface Effects
Marc C. Bétournay
Committee/Advisory Services Project 600511
Technology Transfer to the U.S. Interstate Technical Group on Abandoned Underground Mines
Report MMSL 02-021(TR)
Surface subsidence from underground mining activity is a problem in most countries with underground extraction activity. The impact from such occurrences is well known in coal, metal and other types of mining.
Effective concepts and applications are now being used by related practitioners to identify, measure and mitigate the effect of underground mining. This article provides the reader with a simple overview of the problem and with engineering information on why coal and metal mine extraction openings fail and subsequently impact surface. It defines where and how extensive the problem can be and how to quantify it in order to better understand the nature and magnitude of displacements that can affect surface infrastructure.
In particular, the following aspects are emphasized: initiation of the failure mechanism at the underground opening, the nature of the rock mass displacements to surface, the geomechanical and mathematical expression of the affected surface, the timing of the displacements and the engineering analytical tools that are used to provide the empirical and quantification of these aspects.
2.1.1 Development of Subsidence with Longwall Mining2.2 Metal Ore
2.1.2 Development of Subsidence in Room and Pillar Mining
2.1.3 Time Aspects
2.2.1 Development of Subsidence
2.2.2 Time Aspects
Table 1 - Observed values for the longwall subsidence factor 'a' according to Bräuner
Table 2 - Table of data for adjusting flat seam tensile strains to steep seams
Table 3 - Subsidence duration reported in various countries
Table 4 - Summary of instability elements for failures of shallow stopes of hard rock mines
Table 5 - Duration of metal mining shallow stope failures based on case studies
Figure 1 - Schematic developments of mine effects
Figure 2 - Longwall coal face ripper
Figure 3 - Folded coal and other sedimentary strata, Smokey River Coal Mine, Alberta
Figure 4 - Longwall mining method and ground reaction
Figure 5 - Rock mass behavior above and behind longwall mining face support
Figure 6 - Strata movement above a longwall panel
Figure 7 - Longwall mining subsidence effect 25
Figure 8 - Generic ground and rock mass movements associated with longwall mining
Figure 9 - Subsidence from longwall mining in terms of width and depth of working
Figure 10 - Surface subsidence profiles for dipping seams
Figure 11 - Location of arching over longwall mining
Figure 12 - Subcritical, critical and supercritical subsidence surface effects
Figure 13 - Subsidence surface profile with multiple seam situation
Figure 14 - Subsidence factor versus mining depth in the Appalachian coalfields
Figure 15 - Mathematical form of longwall mining subsidence profile
Figure 16 - Generic strata movement over inclined longwall extraction
Figure 17 - Points of rotation of subsiding ground over dipping longwall mining extraction
Figure 18 - Strains and displacements associated with inclined longwall mining
Figure 19 - Inclined surface subsidence
Figure 20 - Subsidence effects on mining seams close to an inclined surface
Figure 21 - Narrow roadway roof stability conditions
Figure 22 - Common forms of subsidence in room and pillar coal mining
Figure 23 - Surface troughs over room and pillar mining, 0.025 m contour interval
Figure 24 - Chimney caving height calculations
Figure 25 - Time aspects of subsidence over longwall mining
Figure 26 - Common rock mass environments for metal mines
Figure 27 - Common metal mining methods
Figure 28 - Common failure mechanisms of hard rock mines
Figure 29 - Distinct element simulation of block caving
Figure 30 - Ellipsoid drawing pattern for caved rock
Surface impact from mine-derived activity is well known; it can take many forms, and is reasonably well understood. Over the last hundred years or so, effective concepts and theories have been derived with the intensification of coal mining in Europe, and was subsequently complemented with the experience from North America and elsewhere around the world. However, it is only in the last 20 years that metal mining effects have been thoroughly addressed in North America.
Figure 1 provides a generic long-term view of the behavior of rock masses hosting near-surface extracted mine areas. For surface effects to be registered as a result of mining operations, some form of underground collapse (in bulk, piecewise or monolithically) must precede such an effect. Therefore, the form of mining, degree and extent of activity coupled with geological factors play a role in this cause and effect action.
While various mining activities and impacts can occur at various times in the evolution of an undermined site, the long-term view that all underground openings will fail with time is a reasonable assumption (however, there are still numerous examples of underground mine openings that have remained in place for hundreds to thousands of years). This strongly puts into question the consideration we apply to mine opening stability: The location of the openings is forgotten and the relative stability conditions at closure is an indication of long-term safety. Impact assessment and engineering approaches to mitigate short to long-term effects to surface elements must therefore be undertaken.
The objective of this article is to provide the reader with a simple overview and engineering information on why and how coal and metal mine extraction openings fail in North America and subsequently impact surface. It defines where and how extensive the problem can be and how to quantify it in order to better understand the nature and magnitude of displacements that can affect surface infrastructure. More site detailed engineering aspects can be found in other references such as Whittaker and Reddish (1989), who emphasize British coal mining experience, Peng (1992), who focuses on U.S. coal mining experience and Bétournay (2001) who encapsulates mainly the hard rock mining experience.
Coal is a mineral that occurs as a result of accumulation and preservation of organic plant matter. Interlayered in a depositional environment with inorganic sediments, the coal is found in strata interbedded with sandstone, siltstone and shale (decreasing grain size rock types) (Figure 2). The solidification of the organic matter into the solid coal material occurs over time from pressure and heat provided by overlying strata and/or stress from tectonic activity.
Coal beds normally range in thickness from thin layers to several meters. The regional distribution of coal is variable, up to several thousand square kilometers. The common occurrence of coal as near-flat-lying beds in basins has facilitated continuous extraction of a horizon (multiple seams can be found in vertical proximity), and has allowed the application of mechanization and basic ground control as an economic means of extraction. However, many coal fields are folded and faulted (Figure 3), introducing several economic and ground control challenges.
The longwall mining system (Figures 2 and 4), has been commonly used in Europe, the United States and Canada. This system is used to recover the greatest amount of coal in a single, continuous operation, taking advantage of a weak roof which can cave behind the advancing front, thereby minimizing support for temporary needs at the excavation face area.
Preparation work is carried out in parallel transport drifts, fully supported, which are used for transporting the ore and for access to the working area. Breaking the coal at the face is more often done with a drum shearer which sweeps across and shears the coal face from drift to drift. The operation simultaneously loads the broken coal onto a conveyor by which it is moved to the transport drifts.
Only the required space at the coal face is supported. The advance requires the removal of floor to ceiling support furthest behind the face and continued support/advancement at the working face. The procedure is repeated until the entire block is mined. In a true longwall system, the mined-out roof is allowed to cave to the floor, seen in Figures 5 and 6. If it is necessary to minimize subsidence, the space from which the coal has been removed can be filled with some kind of waste material. U.S. longwall mining methods induce complete caving in the mined space [Peng, 1992].
Various degrees of caving movement are encountered from the mined seam, to surface. The mechanism of caving includes detaching of the immediate roof rock into blocks which move irregularly down in the void but results in bulking of that portion of the rock mass of the order of 10-50%. This caved material acts as a natural bed of backfill against which the main roof deflects, thereby restraining its displacement and maintaining its mechanical integrity. Support of loads in this mining system are mobilized in the rock mass by large-scale displacement of pseudo-continuous country rock. Therefore, although gravity potential energy is dissipated by restrained displacement of the country rock, the method also results in the accumulation of strain energy in the near-field rock. It falls between the extremes of fully supported and fully caving methods of mining.
Above this zone, the fractured zone retains a bedding orientation similar to its original geology, as opposed to block caving in hard rock mines which includes block rotation and significant interblock movements, as will be discussed in the hard rock mining section. Between the fractured zone and the surface, strata deform without breaking.
The subsidence zone over this mined out area will generate a lowering of the original surface (Figures 7-8). It can be thought of as a surface wave traveling along the ground surface in the direction of mining advance. A point on surface deforms with an initial tensile phase followed by a compressive phase, and a return to a near-zero strain rate. Structures on surface experience a period of tension, followed by a period of compression but are left in a state of residual tension. Within any given phase, both horizontal movements and vertical (or curvature) movements are expected.
Slip at the soil-structure interface occurs, unless the structure has low tensile strength. As the ground curvature increases, cracking in the structure will occur when the curvature capacity of the footing/base is exceeded. Horizontal movements during the final phase can induce substantial tension and widening of existing cracks. With respect to vertical deformation, the relaxation returns the ground to a near-zero curvature [Lin et al., 1995] once the front is further away.
While the greatest amount of subsidence occurs over the center of the mined-out block, the trough extends beyond the vertical projection of the limits of the block. It is circular to rounded-rectangular, for a square to rectangular mined-opening. The angle between the vertical line through the limit of the mining area and a line connecting the outermost point of surface cracking or subsidence is called the angle of draw. It is 5° - 45° depending on the type of roof material. The surface margin of draw beyond the vertical limit will be 1/4 of the extraction depth for weaker beds, 1/10 for the harder beds [Whittaker and Reddish, 1989].
In the center of the zone of subsidence, the vertical distance by which the original surface subsides has been measured to be as much as half the thickness of the mined coal seam [Stout, 1975], and more [Orchard, 1956-57] (Figure 9). Over dipping coal seams, the surface subsidence is not centered directly over the mined-out room but is shifted slightly down-dip, and the area of subsidence extends farther in that direction (Figure 10).
The location of a pressure arch outline over the opening will determine the extent of surface subsidence. If it remains below surface (Figures 11-12) (width to depth of opening ratio of w/h1.4, "subcritical") versus an arch outline intercepting surface (w/h1.4 "supercritical"), it will discourage maximum surface subsidence. When w/h~1.4, a limit "critical" situation exists. These ratios are used in British coal measures [Whittaker and Reddish, 1989]. Peng  reports that a width and a length times depth of (0.9-2.2)h as a maximum subsidence situation for U.S. mines.
By placing an intervening rib pillar between extraction areas of the same seam, the total subsidence amount can be reduced from one large extraction width. When mining from more than one seam in a vertical succession, ground control aspects (mining the upper seam first will avoid extracting in a disturbed rock, the case if the lower seam was extracted first), and subsidence aspects must be addressed. Seen in Figure 13, the staggering of rib pillars and openings within the same vertical succession allows for less irregular surface subsidence profiles and lesser total subsidence. However, multi-seam interactions are possible [Haycocks et al., 1990] when the depth to seam to distance between the seam exceeds 10:1[Whittaker and Reddish, 1989]. One particular case of folded multiple beds located near surface (<40 m) generate combined effects when the interburden thickness was less than 3 m [Manca et al, 1999]. With high overburden, this imposed pressure includes stress failures within the sill pillar between the horizons when weak units are present, and load transfer from the higher pillars to the lower pillars and floors causing pillar overloading and floor heave. In either case, an increase in anticipated subsidence from a closure of mine openings is possible. [Maletic et al., 1987][Wu et al., 1987].
Field observations indicate that a maximum value of surface subsidence is about 90% of the initial coal seam height, (Figure 9) [Orchard, 1956-57]. Accurate prediction of each of the surface subsidence expression (profile) components of ground movement and deformation (curvature, tilt, horizontal displacement and strain) plays a key role in the estimation of subsidence damage. There are a variety of methods available for the prediction of subsidence. Relatively few of them, however, possess the capability of accommodating the variations of geology and mining conditions or can be used without prior subsidence knowledge. If the calculations are satisfactory in this regard, the parameters are applied to other similar/localized case studies.
The methods of subsidence prediction can be divided into three categories: Profile function (most of the empirical methods), influence function (semi-empirical), and theoretical.
Empirically, the magnitude of the full potential subsidence (critical subsidence) can be expressed as a linear function of the extracted thickness, e.g.
(1) Smax = aM cos
where Smax = maximum value of subsidence
a = subsidence factor
M = extracted seam height
= seam dip
The subsidence factor is dependent on the lithologies and the treatment of caving (total or over some form of stowing), Table 1, Figure 14. It decreases with depth of mining. The severity of surface deformation and resultant discontinuities 'a' is inversely proportional to mining depth.
Liu (1981) and Luo (1989) have quantified 'a' by relating it to rock properties. Peng and Chiang (1984) have provided the following values: for rock of considerable combined strata hardness, a=0.45-0.6; for medium combined hardness, a=0.6-0.8; for soft combined hardness, a=0.8-1.0.
Additionally, the maximum slope of the subsidence trough can be determined from
(2) G = f(Smax/h)
where G = maximum subsidence profile slope
h = depth below surface
The reproduction of the subsidence profile has had many mathematical forms (Figure 15), in the critical case expressed as
(3) s = f(Smax,x, R)
where s = value of subsidence at any point x along the profile
R = critical radius of extraction
An additional parameter, w/h, would have to be considered if the condition is subcritical.
The method is generally applied by adjusting the constants in the equation until a satisfactory shape is obtained which provides an acceptable match to the observed data.
Subsidence parameters and prediction models from case study data have been derived [Peng et al., 1995][Cleaver et al., 1995].
Profile functions predict profiles in set directions across the excavation. They generally use the excavation geometry and tables of data or equations (which relate to mining geometry) derived empirically from coal fields around the world to compare to local observations ('curve fitting') an example of which can be found in Peng and Chen (1981).
Influence functions apply the law of superposition to determine the overall influence (assumed to be circular) of an extraction area treated in infinitesimal parts which are summed to provide the resulting overall subsidence. The functions commonly used can be subdivided into empirical (using arbitrary constants to produce the desired subsidence value and form, and those treating rock properties and failure mechanics. A major advantage of these functions is their ability to predict subsidence movement at any point, but are difficult to apply, check and calibrate.
Analytical models (theoretical, stochastic, and numerical), and physical models applied on a large laboratory scale are also used.
Numerical modelling, while maintaining the advantage of providing detailed surface displacement and strain results, from a global consideration of mining geometry, extraction sequencing, subsurface geology and material behavior, is highly dependent on material property input for result accuracy. The inherent variability and complexity in stratified geomechanical behavior of a rock mass (even if the stratigraphy was well known), and the extremely variable properties of the failed material will engender great discrepancies between prediction and actual field curves.
Given the current computing speed and large scale capabilities, finite element numerical modelling can be used for parametric analyses (targeting actual ground response) with non-lineararity and yield /failure zones incorporated with time steps, in order to obtain a model which can provide accurate results.
Effective application of numerical modelling in coal mining has predominantly been to evaluate the stability of immediate opening periphery, pillars, roofs and floors [Tang and Peng, 1990][Maleki, 1990][Craig et al., 1994][Chandrashekar and Chugh, 1994][Smith et al., 1995][Maleki, et al., 1995][Kripakov et al., 1995]. Attempts at simplifying parameters and using bulk parameters has generated approximate subsidence values and distribution even in complex seam disposition [Manca, 1999].
Some examples of early finite element modelling using linear [Choi and Dahl, 1981][Siriwardane, 1985] and non-linear [Fitzpatrick et al., 1989][Su, 1991] constitutive models have underlined the need for application of anisotropic and non-linear modelling to arrive at results closely agreeing with other subsidence quantification methods and reflecting actual material mechanical behavior such as interbed shearing and separation.
The general character of mining subsidence over steep seams takes on ground movement patterns which are defined by through displacement down-dip from 0°- 45° and trending more to vertical from 45°- 90°, (Figure 10). Figure 16 shows a generic strata movement representation. Results from scale model studies [Degirmenci et al, 1988] show patterns of ground movement which demonstrate that there is a point down-dip side of the extraction around which the subsiding ground tries to rotate, Figure 17. The displacement vector orientation indicate that the surface will experience its greatest horizontal tensile strains on the dip side of the extraction, Figure 18. The rise side may also have, of lesser significance, a point of rotation. Table 2 can be used as a guide to obtain tensile strain values on surface, as corrections to horizontal surface strains.
Studied by stochastic models [Whittaker and Reddish, 1989], the main effect of increasing the surface ground slope angle is to greatly increase the zone of tension of the up-slope side of the extraction. There is also a steepening of the slope locally, Figure 19. There would appear to be conditions which favor increased movement down the slope owing to the significant zone of tensile ground strain created on the up-slope side. The down-slope side experiences increased magnitude of tensile strain effects. The zone of influence is less however, and the rate of change of induced slope due to subsidence is tending to favor increased slope stability with reduced ground slope angle. Smaller radius slopes appear on the up-dip side of the opening compared to downdip slopes.
Influence functions can be used to calculate such surface displacement values [Whittaker and Reddish, 1989].
When sloping ground surfaces and shallow dipping seams are encountered, these will more frequently outcrop. This situation will:
a) lead to weathered rock over the coal seam, increasing sink-hole problems, seen in a clustering of sinkholes close to the outcrop
b) allow the inflow of water, and facilitate depth of weathering, weakening of roof strata so that junctions will more readily collapse, still in a vertical direction.
Extraction from the outcrop of the seam in retreat fashion leads to an initial overburden collapse mechanism where shearing occurs along angle of draw shear lines. Further mining into the hillside leads to a more gradual, typical longwall subsidence pattern. Leaving a coal pillar in the outcrop area and mining in retreat from the slope or parallel to it leaves a subsidence trough and avoids the initial ground shearing. Both are shown in Figure 20.
Narrow excavations such as a roadway have a limited stress redistribution and limited to a collapse process in the area of the roadway, if enough rock mass cover exists. If the pressure within the overlying arch is sufficient to surpass rock strength, fracturing and ground fall will occur, unless support is applied (Figure 21). Artificial support is temporary (timber rots, steel corrodes), and failure may occur in time.
The full vertical extent of failure depends on four conditions: the width of the opening, the propensity of the beds to fail and fall, the bulking factor of the geological units, and if the failed material is removed. The larger the propensity for the material to bulk, such as resistant sandstones, the more difficult is the propensity to initiate and develop failure and the smaller the height of the failure, and vice-versa for very weak rocks such as mudstones. The inherent strengths, block size and bulking factor depend respectively on the ability of the immediate roof to resist tensile and shear stresses, and the inherent weaknesses in the rock such as bedding and cross-joints.
Very common historically, this mining method is suitable and economic in shallow coal deposits (<300 m) with strong immediate roof rock and variations in coal seam continuity/quality. Room and pillar mining has for it's objective to transfer the vertical overburden load to pillars, while the roof may be artificially supported or free standing for the duration of extraction. Some pillars can be recovered in a retreat fashion before extraction is completed thereby leaving a condition for the remaining ones to fail under increased load. Since extraction is so close to surface, it is easy for subsidence features to reach surface, during and many years after extraction. These effects depend on lithological setting, related geomechanical features and mine design.
Two common forms of surface subsidence arise, showed in Figure 22. Subsidence holes and troughs are the main forms of surface instability. The former is defined by the creation of sharply delineated surface sinkholes (cave-ins) formed as a result of upward progressing chimney-shaped failures from mine junctions. These junctions represent the largest mine roof spans). Such failures are common when the roof of the junctions between roadways, in quadrilateral plans (squares, rectangles and rhombohedron) fail to surface. Such a pillar arrangement facilitates operation mechanization. Staggering pillars to form "T" junctions improves roof support capability. Irregular pillar distribution due to variation in ore quality and thickness, leads to troughs and subsidence hole type failures owing to larger roof spans and pillar failure.
The second form of roof and pillar induced subsidence resembles saucer shaped troughs (mostly having disc-shaped profiles and usually not more than 1 m at its center, extensive in area, up to 300 m, depending on the extent of underground pillar crushing) as a result of multiple pillar failures, as seen in Figure 23. It is common with wide room layouts, where narrow rib pillars, for the operation to remain economical, will probably fail with time leading to even wider roof failures.
Underdesigning the pillars in terms of strength can lead to gradual or sudden surface subsidence. In both cases these are time-dependent developments which normally occur after the pillars have sufficiently crushed and that the roof control elements have failed. Pillar punching into mine floors also result in trough subsidence.
Flooding of coal mines after abandonment can deteriorate clay bearing lithologies (e.g. mudstones) which can lead to roof and floor deterioration thereby facilitating closure and rock mass displacements and lead to trough subsidence.
In roof-related stability considerations, good practice will include that a thickness of the immediate roof or well designed artificial support be in place to prevent initial failure. Since slabbing is often the initiating failure due to tensile stresses overcoming rock strength, conventional elastic beam and plate formulas or those developed to take into consideration the difference in rock tension and compression modulus of elasticity [Bétournay, 1995] can be applied.
The failure process of the collapse chimney in weak rock materials [Bétournay, 1998] is an upward process which will be halted only by a strong overlying lithology or by choking of the failure. In weak rocks, the simplified correlation to evaluate whether such a condition will develop (Fs=1) is
(4) Lcritical = 100cm
where Lcritical = maximum stope span for a given rock mass cohesion (metres)
cm = rock mass cohesion (MPa)
Bétournay (1998) derived a limit equilibrium equation to obtain a specific factor of safety against the initiation of chimneying.
Calculation of the height of a chimney before choking occurs takes into account the relevant stages of development of the caved roof strata, Figure 24. The first failure will form a pile with an angle (angle of repose of caved rock). On reaching the roof level, the caved material fills the remainder of the chimney void because of bulking. Equating the total volume of collapsed material with the space available will provide the maximum height of collapse.
(5) z = [4/(k-1)D2](2wM2cot+ Mw2)
where k = bulking factor (lower k, less competent units; coal mine lithologies 1.33 to 1.5)
z = height of chimney
D= chimney diameter
w= width of extraction
M= excavated height
= angle of repose of caved material
The equations assume rock density, material properties and bulking factor does not improve. Whittaker and Reddish (1989) have observed that w<D<w.
The process of chimneying, provided the rock units are weak enough to allow continued development is self-driving. While choking can occur, if the material is removed from the failed pile, the failure will continue. Observations are that in coal mine rock masses and in weak rock masses associated with metal mining, the chimney walls are continuous and roughly parallel.
Various authors [Bruhn et al, 1978][Mahar and Marino, 1982][Marino and Mahar, 1989][Whittaker and Reddish, 1989] have observed that sinkholes that are created to surface are less than 3 m in diameter but can reach up to 10 m, and occur at depths of less than 50 m. While most develop within 50 years of mining, some are reported as occurring 100 years later.
While the process of roof subsidence is relatively simple, the number and variability of the following controlling factors make it difficult to predict with accuracy the time frame involved in subsidence in a case by case basis.
- strength and dimensions of support pillars
- variation in lithological strength properties
- bulking and settlement factors
- water inflow effects
- deliberate caving
- rate of mining
- fill support
- distance from working face
- depth to the seam, seam height and excavation width.
- multiple seam mining effects
Generally the deeper the workings, the longer the time will be for complete subsidence to be registered. In various cases, multiple years (5-8) are required for deep workings (>250 m), and a few years for depths <250 m [Whittaker and Reddish, 1989].
The generic subsidence time curve would include an early portion defining the immediate and most important effects which appear shortly after the cutting operation. The general and physical behavior of the lithologies is one of bed detachment and break-up along jointing and across beds. The dominant effects once the extraction point has passed beyond the critical subsidence distance, is related to residual subsidence effects due to bed closure and consolidation, leading to the flattening out of the time subsidence curve, Figure 25. The initial mine closure and related subsidence is relatively rapid (almost instantaneous at shallow depths, very soon after extraction has occurred (subsidence velocity reaches 2 in/month). The active period sees a subsidence velocity of more than 2 in/month. The residual effects (<2in/month), 5-10% of the maximum subsidence, take much longer to fully develop, up to 12 months and rarely to a few years after the longwall face advancement halts Table 3. The rock mass becomes re-established and there is no risk of subsequent major movements.
Metal-based mineral deposits originate in many ways. Various modes of formation such as depositional, magmatic, intrusive, hydrothermal, epithermal, and enrichment are associated with basic geologic origins: sedimentary, igneous and metamorphic. Overprinted on these origins are the effects of geological processes such as metamorphism, folding, faulting, and alteration which has provided a number of geomechanical rock mass environments in which mining takes place, seen in Figure 26 [Bétournay, 1995].
A variety of mining methods has been applied to metal mining depending on orebody size, orientation and grade, Figure 27.
Because metalliferous orebodies often extend to the upper limit of bedrock, shallow underground stopes have been routinely created close to, or through to, surface. In some cases, mainly related to sedimentary and metamorphic formations, ore horizons do not extend to surface but contain extensive rock mass weakness that could generate failures to surface.
In the upper reaches of bedrock, geological terrains are poor and have variable rock mass quality. Commonly, the first 5 to 15 m of bedrock can be moderately to seriously altered with higher jointing density and wider apertures. Weak rock units formed from geological processes such as faulting, shearing and metamorphism can exist from surface to great depths and therefore represent the initial area and path for certain types of failures.
The geotechnical maxim "each case is a unique case" is applicable here due to the variation in site conditions and rock mechanics parameters. But, as will be seen later, common failure types exist. For the most part, applied ground control to address gravity type failures near surface is limited to mechanical anchors, rarely is backfill used. Sometimes no support at all is applied. Very weak rock masses are given more attention but support for these conditions is sometimes under designed.
Shallow stopes are defined as the underground metal mine openings closest to surface, usually within 30 m, situated at or very near overburden, bodies of water, or infrastructure. The natural tectonic ground stresses may not be sufficient to prevent gravity failures, or in the case of intense regional mining activity, the rock mass may be destressed [Bétournay et al., 2002].
When the rock mass is poorly jointed, and of high quality, extensive lateral stope dimensions, or very thin pillars can be stable in the short term. Movements of the rock mass at the periphery of shallow stopes can be sudden and massive, piecewise and continuous, or gradual over long periods of time, and may not ever reach surface because of re-equilibration, lack of space for the failure material to enter (preventing the failure to continue) or changes in geology or stope configuration. When rock material has poor self-support capabilities, failure to surface by exceeding rock strength is possible. Rock mass instabilities fundamentally originate with geological discontinuities (joints, faults) and rock fabric (bedding, foliation, etc.).
A review of some 110 case studies [Allen, 1934][Bétournay et al., 1987][Bétournay and Labrie, 1988][Bétournay, 1994][Bétournay, 1995][Bétournay and Wang, 1997][CANMET Contract, 1984][CANMET Contract, 1985][CANMET Contract, 1986][CANMET Contract, 1987][CANMET Contract, 1988][CANMET Contract, 1990][CANMET Contract, 1991][Charette and Bétournay, 1992][Charette and Hamel, 1993][Commission d'Enquète Mine Belmoral, 1981][Crane, 1929][Heldley et al., 1979][Rice, 1934][Stefen Robertson and Kirsten, 1984][Wang et al., 1995][Whittaker and Reddish, 1989] indicates that several common metal mine rock mass environments can be identified, in which a possible numbers of failure mechanisms can occur (Figure 28). These include the following geological conditions forming the hangingwall, orebody and footwall: Integral rock material, relatively massive; blocky; well-developed stratification; foliated; and weak, highly fissured, faulted and altered, or strongly schistose, commonly with little self-support capacity.
It has been shown that failures of shallow stopes will not always begin within the surface crown pillar (a mine structure of variable geometry situated between an uppermost stope of a mine and surface) above the shallow stope [Bétournay, 1995]. It could start from the other periphery areas, such as hangingwall or footwall (Table 4). However, most of the time, the failure paths run close to or in the surface crown pillars. Rock mass behavior is therefore based on the portion of the rock mass mobilized in the failure, rather than the thickness of rock available above the stope. Failures reaching surface can begin as deep as 660 m [Allen, 1934]. This imposes a failure mechanism/path focus rather than conventional approaches which rely only on the surface crown pillar, e.g. thickness to width ratio, which can falsely lead to an assumption that a thicker rock mass is safer [Bétournay et al., 1988].
There are six types of common failure mechanisms of shallow stopes of hard rock mines:
Rupture of the surface crown pillar and collapse into the shallow stope.
Sudden fall of the surface crown pillar, delineated by its boundary planes, by gravity into the shallow stope. The planes are well-defined, with near to vertically dipping uninterrupted discontinuities with low shear strength. The pillar thickness can vary; plugs up to 660 m high have occurred [Allen, 1934].
Failure potential depends considerably on the confinement available from redistributed stresses around the stope to resist movement; areas of numerous regional shallow stopes are prime candidates for plug failures.
Rarely do visual precursor movements occur, although smaller scale movements do open discontinuities and allow for water flow. Failure occurs suddenly and completely as a block.
Peripheral block by block rock mass failure without a self-support cavity form reached which includes delamination), ravelling failure involving sliding or buckling of thin rock layers at stope boundaries leading to the deStabilization of the surface crown pillar.
Gradual failure from an unsupported periphery of unfavorably oriented rock blocks, or those where discontinuity orientations are relatively shallow, is commonplace when the span exceeds self-support capabilities, but very low peripheral compressive stresses are required to keep blocks in place [Bétournay, 1995].
Blocks fail without the remainder of the rock mass mobilizing on a large scale unless a stable self-supported arch cavity is formed.
Tensile failure of stratified rock at wall contacts or along the surface crown pillar span.
Few case studies of vertically progressing delaminations have reached surface. The poor cross-jointing normally shown by the rock masses make the strata stiffer and limits failure heights to shallow domes. In other cases, massive block failures have occurred when pillars fail in room and pillar mining. Such failures quickly choke off because of high bulking factors.
The upward progression of disintegration within a weak rock mass forming a cavity with limited lateral extent (<5 m), developing from an underground opening toward surface. This condition is similar to that described for roadway intersections in room and pillar coal mining.
Failures (self-driving, and continuous if failed material is removed) normally progress upward in thick weak geological units (with low cohesion that can be sheared easily), but will continue up-dip when stronger units are encountered [Bétournay, 1998]. Equations 4 and 5 can be used respectively to evaluate the possible initiation of the failure mechanism and its maximum height. The bulking factor for weak rock units which host this failure mechanism in hard rock mines (e.g.schists, altered rock) is 1.05<k<1.2.
Ground support is only marginally effective in providing anchoring and global peripheral integrity.
rock mass caving: Rock mass break-up and gravity mobilization of blocks (flow) towards and into an opening leading to a progressive failure front moving towards surface.
Although the initiation of rock mass caving can begin with block ravelling, the mechanical action involved in breaking and mobilizing the rock mass (Figure 29) is difficult to quantify or predict with accuracy. Rock masses with blocks of similar shapes, smaller sizes with low surface friction, and shallow angle discontinuities are likely in areas of low confining stresses to block cave. These and many other factors readily make it difficult for a rock mass to breakdown during and after mining extraction.
The dimensions of the problem, along with surface effects have been defined by Janelid and Kvapil (1966) based on silo studies. The volume that has caved into a stope is defined by the draw ellipsoid. The limit ellipsoid contains the zone of broken material that has subsequently moved and expanded under gravity to fill this volume. If the draw ellipsoid intersects surface, complete failure of the rock mass in the surface crown pillar is registered (Figure 30).
Some authors refer to several types of failure by the generic name "chimneying" because of the geometry of the failure boundary. Plug failure, chimneying in weak rock masses and piping through caved material are described as chimneying caving. The nomenclature of the failures given here is based on their mechanical behavior to avoid confusion stemming from shape and to allow for a more scientific approach, analysis and communication.
The effects of metal mining rock mass movements on surface is unlike the broad regional subsidence related to longwall mining or troughs that exist over failed areas of pillars in coal mining extraction described earlier. In those circumstances, the bedrock surface and the overlying soil is lowered. More often than not, a smooth dipping surface results , with no abrupt changes in elevation.
Cave-ins, discontinuous subsidence, are the norm in metal mines, with voids or partially filled voids connecting the surface with underground openings. In most cases (except block caving), the surface does not significantly subside until the last portions of the upward developing failure reach or are about to reach the top of bedrock. Therefore, the limit equilibrium equations to quantify the factor of safety [Bétournay, 1995] and an empirical relationship for dividing failed and unfailed cases [CANMET Contract, 1990] will not provide a subsidence quantification or profile of the lowered surface.
Owing to the relatively limited number of geomechanical variables related to metal mine site rock masses, numerical modelling provides the possibility of obtaining approximate solutions to the behavior of individual or adjoining and even regional shallow stopes, while considering all of the influencing factors. The use of appropriate modelling approach can simulate material behavior (elastic, non-elastic, discontinuous, time-dependent) and the geometry of the case, in 2D or 3D.
The discontinuous block codes will be best suited to discontinuous rock masses, reflecting block ravelling, plug, destratification and block caving failures accurately. Elastic finite element codes will be best suited to massive rock masses for evaluating rock fracturing, and non-linear versions to evaluate failures in weak rock masses normally subject to chimney disintegration.
In both numerical approaches, final surface profile and strain values are provided. In certain non-elastic finite element modelling codes, time steps are available to see the progress of the failure/surface profile with time.
Numerous case studies of failure mechanism based numerical modelling of shallow stopes have been reported [Bétournay et al., 1987][Mirza et al., 1989][Bétournay, 1995][Bétournay and Mitri, 1995][Wang et al., 1995].
According to the general observations of failed case studies, once failure is initiated, and changes in stope geometry or rock mass conditions impose no effect, most failure mechanisms will develop in relatively short time spans (Table 5) (should any ground support applied not be effective and stope space available to allow for its continuation). Although it is necessary that evaluations be carried out on simple and complex failure development scenarios, the simple influencing factors should be kept in mind.
Carter and Miller , considering a 230 case study used to derive a dedicated empirical stability evaluation [CANMET Contract, 1990], made some broad observations on time dependent failure considerations. A bi-modal occurrence of failure has been observed. Initially, after mine closure (about 20 years), a maxima of failed cases occurs (reaching an ~3% failure rate) which the authors ascribe to failures based on rock mass "effects", while a second peak of ~2.5% failures is reached in the 60-year timeframe, reflecting "wearing out" effects. Very little research has been performed on long-term discontinuity shear strength which would have a bearing on the weakening of the interblock resistance.
What is evident is that with pre-existing rock mass conditions that will not rapidly lead to failures (e.g. two joint families only where three is needed for block development, poorly intersecting joint relationships) which can be thought of as higher quality rock masses at one end of the classification spectrum, it is anticipated that such a shallow stope setting will not fail until a much later time frame. At the other end of the spectrum, rock masses with poor configurations, or weak, will be expected to fail.
Certain rock types are prone to decomposition or to disintegration. Progressive failure in this environment has an advance rate of about 5 m per 50 years.
Allen, C.W., 1934. Subsidence resulting from the Athens system of mining at Negaunee, Michigan, Proceedings Am. Inst. Min. & Met. Eng.,109, pp. 195-202.
Atlas Copco, 1973. A brief guide to mining methods, Atlas Copco, 32p.
Bétournay, M.C., Yu, Y.S. and Thivierge, S., 1987. A case study of surface crown pillars: the Niobec Mines, Proceedings 28th U.S. Rock Mechanics Symposium, Tucson, Balkema, pp. 1197-1204.
Bétournay, M.C. and Labrie, D., 1988. La stabilité des chantiers supérieurs et leurs piliers de surface, Mine Eldrich: méthodes analytiques, CANMET Division Report MRL 88-17(TR), Energy, Mines and Resources, Canada, 45p.
Bétournay, M.C., Nantel, S. and Lessard, D., 1988. Summary of 24 surface crown pillar case studies, CIM Bull., 81 (917), pp. 73-77.
Bétournay, M.C., 1994. Chimneying disintegration failure mechanisms of hard rock mines, Proceedings 1st North American Rock Mechanics Symposium, Austin, Balkema, pp. 987-996.
Bétournay, M.C., 1995. The stability of shallow stopes of hard rock mine, McGill University Ph.D. Thesis, 611p.
Bétournay, M.C. and Mitri, H.S., 1995. Numerical modelling as a tool for stability of shallow stopes of hard rock mines, Proceedings 3rd Canadian Conference on Computer Application in the Mineral Industry, Montreal, pp. 446-455.
Bétournay, M.C. and Wang, B.W., 1997. Review of the impact of mining on the shallow stope rock mass at the Lamefoot Mine, Washington State, Phase II numerical modelling and analysis of rock mass displacements and subsidence, CANMET Report MMSL 97-058 (CR), Natural Resources Canada.
Bétournay, M.C., 2001. A Canadian standard for hard rock mine shallow stope decommissioning, Natural Resources Canada Special Report.
Bétournay, M.C., Boyle, R. and Li, G., 2002. Impact of regional destressing from intense mining on long-term inactive mine stability, Timmins, Ontario, Proceedings, 55th Canadian Geotechnical Conference, Niagara Falls, in print.
Brady, B.G.H. and Brown, E.T., 1985. Rock mechanics for underground mining, George Allen and Unwin, 527p.
Bräuner, G., 1973. Subsidence due to underground mining. Part 1 Theory and practices in predicting surface deformation. Bureau of Mines, U.S. Department of the Interior, 56p.
Bruhn, R.W., Magnuson, M.O. and Gray, R.E., 1978. Subsidence over the mined-out Pittsburgh Coal, A.S.C.E. Convention , Coal Mine Subsidence Session, Pittsburgh.
CANMET Contract 26sq 23440-3-9005, 1984. Surface Pillars, Le Groupe Conseil Roche Ltée.
CANMET Contract 26sq 23440-5-9014, 1985. Surface Pillars Phase II, Le Groupe Conseil Roche Ltée.
CANMET Contract 15sq 23440-5-9017, 1986. Sampling, field testing and modelling of a surface crown pillar at Les Mines Selbaie, Joutel, Québec, Mirza Engineering.
CANMET Contract 01ss 23440-7-9153, 1987. Seismic characterization of discontinuities and anomalous rock quality within mine surface crown pillars using attenuation and velocity imaging, Queen's University.
CANMET Contract 03sq 23440-8-9063, 1988. The determination of surface crown pillar mechanical and structural properties, Trow Engineering.
CANMET Contract 01sq 23440-8-9074, 1990. Crown pillars stability back analysis, Golder Associates.
CANMET Contract 01sq 23440-9-9194, 1991. Stability of the overburden hangingwall, Gays River Mine, Nova Scotia, Westminer Canada Ltd.
Chandrashekhar, K. and Chugh, Y.P., 1994. Three dimensional modeling of a four-way coal mine intersection for analysis of roof-pillar-floor interaction, Proceedings 1st North American Rock Mechanics Symposium, Austin, pp. 1007-1016.
Charette, F. and Bétournay, M.C.,1992. Évaluation géomecanique préliminaire des piliers de surface à la mine Lamaque, CANMET Report MRL 92-04, Energy Mines and Resources Canada.
Charette, F. and Hamel, G., 1993. Projet d'évaluation de la stabilité des excavations et des piliers de surface des lentilles A-1, 2, et 3, Mines Selbaie, CANMET Report MRL 93-059 (CL), Energy, Mines and Resources, Canada.
Choi, D.S. and Dahl, H.D., 1981. Measurement and prediction of mine subsidence over room and pillar in three dimensions, Proceedings of the Workshop on Surface Subsidence Due to Underground Mining, West Virginia University, pp. 34-47.
Cleaver, D.N., Reddish, D.J., Shadboldt, C.H. and Dunham, R.K., 1995. Statistical regional calibration of subsidence prediction models, Proceedings 14th International Conference on Ground Control in Mining, Morgantown, pp.236-243.
Commission d'Enquète sur la tragédie de la mine Belmoral et les conditions de sécurité dans les mines souterraines, 1981. Volume 1, Les Mines Belmoral Ltée, cause et prévisibilité de l'effondrement.
Craig, M.S., Duncan Fama, M.E., and Trueman, R., 1994. Three-dimensional elasto-plastic analysis of coal pillars, Proceedings 1st North American Rock Mechanics Symposium, Austin, pp. 999-1006.
Crane, W.R., 1929. Subsidence and ground movement in the copper and iron mines of the upper peninsula, Michigan, U.S. Bureau of Mines, Bulletin 295.
Degimenci, N., Reddish, D.J. and Whittaker, B.N., 1988. A study of surface subsidence behavior arising from longwall mining of steeply pitching coal seams, Proceedings 6th Coal Congress, Zonguldak, 27p.
Fitzpatrick, D.J., Reddish, D.J., and Whittaker, B.N., 1986. Study of surface and subsurface ground movements due to longwall mining operations, Proceedings 2nd Workshop on Surface Subsidence Due to Underground Mining, Morgantown, pp. 68-77.
Haycocks, C., Holland, C.T. and Zhou, Y., 1990. Multiple-seam mining - a state-of-the-art review, Proceedings 9th International Conference on Ground Control in Mining, Morgantown, pp.1-11.
Hedley, D.G.F., Herget, G., Miles, P. and Yu, Y.S., 1979. Case history of CANMET's rock mechanics research at the Kidd Creek Mine, CANMET Report MRL 79-47 (TR), Energy, Mines and Resources Canada.
Janelid, I. and Kvapil, R., 1966. Sublevel caving, Int. J. Rock Mech. Sci., 3, pp. 129-153.
Kolebaevna, J.M., 1968. Theory of ground movement in rock mass, Nedra (Moscow)(in Russian), 101p.
Kripakov, N.P., Sun, M.C. and Oelfke, S.M., 1995. Overview of recent developments by USBM in coal mine design using numerical modeling techniques, Proceedings 3rd Canadian Conference on Computer Applications in the Mineral Industry, Montreal, pp.197-207.
Liu, T.Q., editor, 1981. Surface movements, overburden failure and its applications, Coal Industry Publishing, China, 369p.
Luo, Y., 1989. An integrated computer model for predicting surface subsidence due to underground coal mining, Ph.D. Thesis, West Virginia University, 168p.
Mahar, J.W. and Marino, G.G., 1982. Building response and mitigation measures for building damage in Illinois. Proceedings of Workshop on Surface Subsidence due to Underground Mining, Morgantown, West Virginia University, pp. 238-252.
Maleki, H.N., 1990. Development of modeling procedures for coal mining stability evaluation, Proceedings 31st U.S. Symposium on rock mechanics, Golden, pp. 85-92.
Maleki, H., McKibbin, R.W., Jones, F.M., and Wheeler, R.A., 1995. Application of numerical modelling to the analysis of strata interactions, Proceedings 3rd Canadian Conference on Computer Applications in the Mineral Industry, Montreal, pp.187-196.
Manca, P.P., Lonis, R., Manca, F. and Sirigu, E., 1999. Prediction and control of surface subsidence over abandoned mined lands in Carbonia, Proceedings 18th International Conference on Ground Control in Mining, Morgantown, pp. 315-321.
Marino, G.G. and Mahar, J.W., 1985. Subsidence damaged homes over room and pillar mines in Illinois, Report for Illinois Abandoned Mine Lands Reclamation Council, 450p.
Matetic, R.J., Chekan, G.J., and Galek, J.A., 1987. Design considerations for multiple-seam mining with case studies of subsidence and pillar load transfer, Proceedings 28th U.S. Symposium on Rock Mechanics, Tucson, pp.1095-1106.
Mirza, C., Lau, K.C., Bétournay, M.C. and Crawford, A., 1989. A complete recovery of surface crown pillars without removal of overburden, Proceedings International Conference on the Evaluation of Surface Crown Pillars for Active and Abandoned Metal Mines, Timmins, pp. 113-122.
National Coal Board, 1975. Subsidence Engineer's Handbook, National Coal Board, 111p.
O'Connor, KM, Chandler, M.R., Miller, S.M., Siekmeir, J.A.and Stache, J.E., 1996. Uncertainty in evaluation of historical subsidence measurements, article submitted to ASCE Conference on Uncertainty in Geological Environment, Madison.
Orchard, R.J., 1956-57. Surface effects of mining - the main factors, Trans. Inst. Min. E., 116 part 1, pp. 942-944.
Orchard, R.J., 1956-57. Surface effects of mining - the main factors, Trans. Inst. Min. E., 116 part 2, pp. 941-955.
Pasamehmetoglu, A.G., 1972, An investigation into time-dependent aspects of mining subsidence, Ph.D. Thesis, University of Nottingham, 190p.
Peng, S.S., 1992. Surface subsidence engineering, Society for Mining, Metallurgy, and Exploration, 161p.
Peng, S.S. and Chen, D.W., 1981. Analysis of surface subsidence parameters due to underground longwall mining in the North Northernlachian coalfield, West Virginia University, TR81-1, 22pp.
Peng, S.S. and Chiang, H.S., 1984. Longwall mining, Wiley, 708p.
Peng, S.S., Luo, Y. and Zhang, Z.M., 1995. Development of a subsidence database and determination of subsidence parameters, Proceedings 14th International Conference on Ground Control in Mining, Morgantown, pp.226-235.
Rice, G.S., 1934. Ground movement from mining in Brier Hill Mine, Norway, Michigan, Proceedings Am. Inst. Min. & Met. Eng., 109; pp. 118-152.
Siriwardane, H.J., 1985. A numerical procedure for prediction of subsidence caused by longwall mining, Proceedings 5th International Conference on Numerical Methods in Geomechanics, Boston, pp. 1595-1602.
Smith, W.C., Dolinar, D.R. and Khamis, Y.H., 1995. Nonlinear approach for determining design criteria for yield pillar performance, Proceedings 3rd Canadian Conference on Computer Applications in the Mineral Industry, Montreal, pp.177-186.
Stefen Robertson and Kirsten, 1984. Rock mechanics study, Thompson open pit, Report for INCO, 55p.
Stout, K.S., 1975. Soil and rock mechanics illustrated, Bulletin 96, Montana College of Mineral Science and Technology, 190p.
Su, D.W.H., 1991. Finite element modeling of subsidence induced by underground coal mining: The influence of material non-linearity and shearing along existing planes of weakness, Proceedings 10th International Conference on Ground Control in Mining, Morgantown, pp. 287-300.
Tang, D.H., Peng, S.S., and Holland, C.T., 1990. Mine stability analysis using FEM methods - two case studies, Proceedings 9th International Conference on Ground Control in Mining, Morgantown, pp. 88-98.
Voegele, M., Fairhurst, c. and Cundall, P., 1978. Analysis of tunnel support loads using a large displacement, distinct block model, in Storage in Excavated Rock Caverns, Pergammon, V2, pp. 247-252.
Wang, B.W., Yu, Y.S. and Aston, T., 1995. Stability assessment of an inactive mine using the
block-spring model, Proceedings 3rd Canadian Conference on Computer Applications in the Mineral Industry, Montreal, pp. 390-399.
Whittaker, B.N. and Reddish, D.J., 1989. Subsidence: Occurrence, prediction and control, Elsivier, Amsterdam, 528p.
Wu, W., Haycocks, C., and Zhou, Y., 1987. Designing for interaction in close-seam multi-seam mining, proceedings 28th U.S. Symposium on Rock Mechanics, Tucson, pp. 1107-1114.
Table 1. Observed values for the longwall subsidence factor 'a' according to Bräuner (1973)
|Coalfield||Form of Goaf Treatment||Subsidence Factor 'a'|
|British coalfields||solid stowing||0.45|
|caving or strip-packing||0.90|
|Ruhr coalfield, Germany||pneumatic stowing||0.45|
|other solid stowing||0.50|
|North coalfields, France||hydraulic stowing||0.25-0.35|
|Upper Silesia, Poland||hydraulic stowing||0.12|
|Russia||caving, Donets and Kizelov coalfields||0.60|
|caving, Kuznetsk and Karaganda coalfields||0.70|
|caving, Chelyabinsk coalfield||0.90|
|Seam Dip (degrees)||Gradient||Proportion of Normal Tensile Strain|
|33.7||1 in 1.5||0.29||1.71|
|26.6||1 in 2.0||0.35||1.65|
|21.8||1 in 2.5||0.41||1.59|
|18.4||1 in 3.0||0.46||1.54|
|15.9||1 in 3.5||0.51||1.49|
|14.0||1 in 4.0||0.56||1.44|
|12.5||1 in 4.5||0.60||1.40|
|11.3||1 in 5.0||0.64||1.36|
|10.3||1 in 5.5||0.69||1.31|
|9.5||1 in 6.0||0.73||1.27|
|8.7||1 in 6.5||0.76||1.24|
|8.1||1 in 7.0||0.80||1.20|
|7.6||1 in 7.5||0.83||1.17|
|7.1||1 in 8.0||0.85||1.15|
|6.7||1 in 8.5||0.87||1.13|
|6.3||1 in 9.0||0.89||1.11|
|6.0||1 in 9.5||0.91||1.09|
|5.7||1 in 10.0||0.92||1.08|
|Country||Mining depth (ft)||Face advancing rate(ft/day)||Total duration (month)||Active period (month)|
|Type of Failure||Propitious Environments ^(Figure 25)||Mobilized Rock Mass||Lack of Ground Stress Clamping||Controlling Instability Elements|
|Rock fracturing||a||FW, CR, HW||No||Stress loading (p) (gravity induced)|
|Plug failure||d, e, f||CR||Yes||Near-vertical continuous discontinuities (p); low friction surface properties (s)|
|Raveling||b, e, f, g||CR, HW||Yes||Blocky rock mass (p)|
|Strata failure||f||CR, HW*||No||Stratification (p); stope span (p)|
|Chimneying disintegration||c, d, e, g||CR, HW||No+||Material of low cohesion (p); small size rock mass pieces (s)|
|Block caving||b, e, g||CR, HW||Yes||Well-developed jointing and blocks (p); stope span (s); similar block size (s)|
+ can be overstressed * shallow dipping order of importance: p primary, s secondary
^a poorly jointed, b jointed and blocky, c schistose orebody/competent walls, d competent orebody, schistose walls, e generally foliated/slaty, f well developed stratification, g fault-weakened rock mass
Table 5. Duration of metal mining shallow stope failures based on case studies [Bétournay, 2001].
|rock fracturing||sudden but full break and fall can be longer term|
|plug failure||instantaneous once initiated|
|destratification||several meters per day|
|chimneying disintegration||one strata per day to several months|
|block caving||twenty to one hundred meters per year|