Catastrophic Failures of Underground Evaporite Mines

Jeff Whyatt, Mining Engineer

Floyd Varley, Acting Branch Chief
NIOSH - Spokane Research Laboratory
Spokane, WA

ABSTRACT acceleration, or at least enough energy to remove any hazard to

miners. That is, the failure process will be sufficiently slow to
Deformation of underground salt, trona and potash mines is give miners time to retreat. However, the economic and
generally time dependent, providing for gradual adjustment of environmental consequences (inundation, sinkholes, etc.) may
strata to mining induced stresses. Time dependence can allow for still be catastrophic. The alternative is sudden, brittle failure.
higher extraction ratios provided eventual failure can be tolerated.
However, this eventual failure can be violent if creep deformation Brittle failures of any appreciable size around working areas
can shift stress and potential energy to strong, brittle geologic can be hazardous. The unexpected transition from a controlled
units. The mine failure case studies reviewed here illustrate this ground failure process (that is, ductile failure and/or brittle failure
process. Yield pillars and defects in bridging strata figure isolated from miners) to a brittle failure process that impacts work
prominently in these cases. Yield pillars provide local and areas is a hallmark of many fatal accidents and mine disasters. At
temporary support to the roof, temporarily delaying the cave; and ordinary working temperature and pressure, salt rocks
allowing extraction ratios and overburden spans to increase (evaporites) are considered to be the most ductile of all rock types
beyond the long term capacity of overlying strata. Defects (faults, (1). As such, mines extracting such rock might be expected to be
voids, thinning) of strong overburden strata reduce the critical immune from such failures.
span, sometimes to less than panel width. Analyses of many of
these cases have focused on a cascading pillar failure mechanism, This paper examines catastrophic failures of evaporite mines
but recent work and this review point to failure of strong (including salt, potash and trona) that show that violent failure
overburden strata as the essential element. The suddenness of does, in fact, occur. It also attempts to identify common factors
failure and attendant seismic events pose hazards to miners and, and characterize the mechanism of failure. While these cases are
in some cases, to those on the surface. Characterizing these limited to evaporite mines, it is hoped that results are sufficiently
failures is a first step towards recognizing and managing the risk general to provide insight into similar cases in other mining
of catastrophic collapse in underground mines. sectors.

INTRODUCTION CASE STUDIES

Recent mine disasters have prompted the National Institute for Cases selected for this review involved panel or mine collapse
Occupational Safety and Health (NIOSH) to look for events that were initiated by strength failure of solid material.
opportunities for reducing the disaster potential of mining. A Events driven primarily by gas (gas outbursts) or groundwater
number of other agencies and groups are addressing the (inundations) in the absence of an initial collapse of strata were
regulatory and management frameworks for controlling disaster not considered. However, many of the cases reviewed do involve
potential. While the details of these schemes differ in assigning release of gas and/or water.
responsibilities and liabilities, all require technical means to
discern and minimize the likelihood of catastrophic mine failure. Varangeville Mine Collapse (1873)

Deep and high extraction mining is possible only through This, the oldest collapse case reviewed, occurred in 1873 at the
management of rock mass failure. Generally, two types of failure Varangeville salt mine. Salt was extracted by blasting and
may occur. First, yielding around mine openings shifts stress solution mining for an ultimate extraction ratio of over 82%,
away from openings towards more confined rock. Second, caving leaving pillars with a width:height ratio of about 1:1 (2).
shifts the weight of caved material from pillars and abutments to Significant subsidence was noted, beginning in October, 1873.
the floor of mined areas. Creation of a fissure in a building at 5 am on October 31st
prompted evacuation shortly before a collapse that reportedly
These types of failure may occur in two modes – ductile and took less than 2 seconds to occur. Fractures were formed within
brittle. Ductile failure consumes sufficient energy to prevent its the subsidence zone in two concentric circles with 160-m and 350
-m diameters, respectively. Daw and Pollard (3) describe geology
elsewhere in the deposit as having a surface stratum of 42 m of
marl underlain by layered salt and marl. More recently, solution
mining in the deposit has been conducted under 200 to 300 m of
cover (4). The thick surface marl forms a brittle cap with
increased calcium carbonate content. Apparently, there was a
transition from ductile to brittle failure as caving entered the marl
which formed a brittle cap with increased calcium carbonate
content.

Louisiana Salt Dome Failures to Surface (1972-1992)

Rapid, but not seismic, collapses have occurred in Louisiana

salt domes. The south coast of Louisiana contains five salt domes
that have been mined since the beginning of the 20th century (5).
More recently, two of the domes have experienced large-scale Figure 1. Idealized cross section through the long axis
failures to the surface as a result of unsustainable ductile failure of the Prairie Evaporite structure (11; 12).
long after mining of pillars in the affected area. The salt domes
are massive deposits with discontinuous jointing and inferred
shear zones overlain by a more brittle "caprock" followed by noise, movement of air (minor air blasts) and falls of loose roof
surface soils. rock have. Hasegawa et al. suggest that salt overlying mining
provides protection by attenuating dynamic shocks.
At the Belle Isle mine, extraction ratios between 50 and 59%
initiated creep in the salt structure with minimal disturbance of Sepehr and Stimpson (10) created a simplified numerical
the active mine. Creep transferred loads to more brittle cap rock, model of strata failure, inspired by a potash mine near Saskatoon,
eventually resulting in formation of a 70 meter diameter sinkhole Sakatchewan that had produced seven events with magnitudes of
in 1972. The sinkhole developed over about 15 minutes and 2.3 to 3.5. The model simulated mining of evaporite beds under 1
engulfed the lone shaft accessing underground workings (5). km of overburden. They found failure propagation into the
30,000 m3 of debris flowed into the mine (3). Fortunately, and by Dawson was “so rapid and extensive… that numerical
chance, no miners were entrapped or injured. The shaft was convergence is not achieved, signaling a structurally unstable
backfilled, a new shaft sunk, and the mine returned to operation. situation.” They concluded “such extensive and rapid brittle
Subsidence monitoring showed increased risk of a second, much failure would certainly induce significant seismicity.” That is,
larger event in 1983. The mine was then flooded to stabilize they interpret failure of the solver algorithm in their finite element
workings and prevent damage to nearby oil and gas production. program as indicating physical as well as numerical instability.

At Weeks Island, parts of the original mine workings, dating to Un-named Phosphate Mine Collapse
the late 19th century, were taken by the U.S. Department of
Energy for use as a petroleum reserve (6). Mining operations Chen and Peng (13) report on a relatively innocuous collapse
were moved to an adjacent part of the dome and monitoring of that occurred two weeks after completion of mining. The collapse
surface subsidence and water inflow initiated. A number of occurred within a pillar retreat panel with strong roof and floor,
production levels were established in the new mine and mining mined under 60 to 400 m of overburden near a cliff. The roof had
progressed without incident for more than a decade. As at Belle remained intact during secondary mining. The collapse occurred
Isle, the brittle cap-rock became over-loaded and a sinkhole suddenly, crushing 25 remnant pillars, all less than 6 m in width
developed at the surface along the boundary between old and new and 1.6 to 1.9 m in height, over a roughly 90 m by 55 m area
mines (7). Increased leakage of surface waters into the reserve (Figure 2). The collapse was also evident in surface subsidence
was detected. The progression of sinkhole and leakage was slow, and large cracks in the cliff face.
but still posed a significant risk to the petroleum
reserve. Grouting failed to stem progression of failure. The Cheng and Peng attributed the failure to “pillar sizes that were
reserve was stabilized by brine flooding and abandoned. too small and uneven.” Implicit in this explanation is a failure of
overlying strata to transmit overburden loading to pillars beyond
Saskatchewan Potash Mining Seismic Events (1962-1989) the collapse area. Fortunately, the pillars succeeded in providing
support during mining of the area.
Some evaporite mines are a significant source of mining
induced seismic events. Hasegawa et al. (8) reviewed large
seismic events produced by mining of potash in Saskatchewan Retsof Salt Mine Collapse (1994)
under 900 to 1000 m of overburden. These deposits lie in a thick
sequence of halite and anhydrite known as the Prairie Evaporite, Major seismic collapse events can also be linked to other
which is overlain by roughly 40 m of strong, dense Dawson Bay mining hazards, before and after the collapse. For instance, a
limestone (Figure 1). chain of events at the Retsof salt mine, New York, began with
ground control problems encountered under roughly 300 m (1000
Between the onset of mining in 1962 and the end of 1989, 21 ft) of overburden, including a roof fall that caused two deaths in
mining induced seismic events with magnitudes of 2.3 to 3.6 were 1990. In response, the yield pillar panel method of mining was
recorded. Horner (9) attributes these events to “brittle failure or tested and then implemented in two full production panels, 2YS
sudden rupture” in the competent carbonate rock of the Dawson and 11YW (Figure 3). Mining of these panels was initiated in
Bay limestone. Rockburst damage has not been reported but 1993 but 2YS was halted in October “for safety reasons and due
 strata – nor by the yield pillars for which such loading was not
anticipated.

Solikamsk-2 Potash Mine Collapse (1995)

Collapse of the Solikamsk-2 potash mine, Verkhnekamsky

deposit, in the Upper Kama district of western Ural, Russia
resulted in a 4.7 magnitude seismic event on January 5th, 1995
and 4.5 m of surface subsidence (17). Underground, a “massive
falling of the mine roof” was noted over a 600 m by 600 m area
(Figure 4). The event released an estimated 900,000 m3 of gas (a
mix of methane, hydrogen, carbon dioxide, carbon monoxide and
other gases). These release led to gas explosions the following
day. Timely placement of a “large volume” of backfill is credited
with preventing further catastrophic consequences. Malevichko et
al. describe the potash as “almost incompressible, highly ductile
and rather easily deformed by creep.” Potash and salt beds were
overlain by carbonates and sandstones.

Figure 2. Collapsed pillars (shaded) and surface cracking

(dashed lines) at a Phosphate mine (13).

Figure 4. Subsidence resulting from collapse of the Solikamsk

-2 Potash Mine, Russia, in 1995 (Modified from 17).

Sylvinite (potash) was extracted from two to three beds with

10 to 16 m of total extraction height in the collapse area. The
Figure 3. Yield pillar panels at the Retsof mine. mine used a panel system of rooms and pillars under 200 to 400
The 2YS panel collapsed on March 12, 1994, leading m of overburden. Rooms were 13-16 m wide and pillars 11-14 m
to inundation and loss of the mine (14). wide by 200 m long.

to clearance problems for the mining equipment” (14). A sudden An investigation across the Upper Kama district found that
increase in closure rates in both panels 2YS and 11YW led to surface subsidence typically reached 50% of excavation height 4 ­
cessation of mining in panel 11YW on March 1, 1994. 8 months after excavation. This event occurred 15 years after
mining began, and 7 years after mining was completed (18). The
Collapse of the 2YS panel on March 12, 1994 produced a 3.6 unusual delay in subsidence implies extensive bridging of
magnitude seismic event as a 150 by 150-m (500 by 500-ft) overburden, consistent with its uncommonly high release of
section of shale roof collapsed. Methane and hydrogen sulfide seismic energy. The next largest collapses occurred in 1993 and
gases were detected and brine water began flowing into the mine 1997 with magnitudes of 2.6 and 2.8, respectively.
at nearly 19,000 lpm (5000 gpm) (15). The collapse stabilized
closure of 11YW panel but flooding could not be stopped, 3rd Bereznki Potash Mine Collapse (1986)
eventually leading to loss of the mine.
Collapse of the nearby 3rd Bereznki potash mine in July 25,
Subsequent investigations found evidence of fracture zones 1986 (19) is another interesting case. Mining was active at depths
and a brine and gas pool 50 m (160 ft) above the mining horizon of 235 and 425 m below surface, and was overlain by a 100 m
(14; 16). The fractures and pool weakened the overlying “bridge” thick “salt complex” of halite and carnallite beds, overlain in turn
of strata. Fractures hydraulically connecting the pool to surface by clays, carbonates, aquifers and sediments. Mining formed
waters allowed recharging of the pool and maintenance of yield pillars with 5.3 m wide rooms, 3.8 m wide pillars and a 5.5
hydrostatic pressure. Since this pressure was exerted within the m mining height. After mining, conditions were described as
bridging arch, it could not be entirely supported by intervening “pillars crushed and roofs sagged.”
 Figure 5. Collapse and explosion of cavity on July 25, 1986 at the 3rd
Bereznki Potash Mine in Russia (redrafted from 19).
At the collapse site, mining induced fractures connected with
natural fractures in a fold structure, providing a conduit for
flooding of the mine. Loss of hydraulic control in January of 1986
led to creation of a large cavity beneath a sandstone/limestone
sequence nearly 200 m thick that was relatively stable until July.
Failure of this sequence began at 18:30 hours with “clearly felt
underground shocks” culminating with a final collapse at
midnight “accompanied by an explosion with flashes of
light” (Figure 5).

Solvay Trona Mine Collapse (1995)

A roughly 1 by 2 km section of the Solvay Mine, Wyoming

collapsed on February 3, 1995, causing a 5.1 magnitude seismic
event (20). Seismic first motion showed dilation (collapse) at all
seismic stations. The “dominant movement of the ground” took
place on a “time scale of a few seconds or less” (21). Miners
described the event as “a rumbling, a big boom, and then a
deafening sound lasting 5 to 6 seconds in all.” The collapse
caused an air blast fed by methane and ammonia emissions as
well as closure. The failure was believed to have released an
eventual total of 3 million m3 of methane from broken shales with
a peak release of nearly 1 million m3 / day. Ammonia (broken
trona and oil shales) and CO (kerogen-rich oil shale in the
immediate floor) were released in smaller quantities. The single
fatal injury was attributed to ammonia poisoning.

The failure occurred over a multiple yield panel section within

the southwest portion of the mine but was contained by barrier
pillars along the main entries (Figure 6). All 13 panels in the
southwest section were “completely collapsed or extensively
caved” (22). Panels 4W through 12W were “caved tight at or near
the submain ends.”

The panel was mined under 450 to 520 m (1500 to 1700 ft) of Figure 6. Schematic of Solvay mine in 1995 showing (A)
overburden, including the massive Tower Sandstone. Failure of southwest panels (1W through 12W) closed by the collapse,
yield pillars within panels occurred in the immediate floor and (B) location of the southwest panels within the mine and (C)
lower portion of narrow pillars, while the immediate roof and detail showing barrier pillars that bounded the collapse
upper portion of pillars remained intact. Gateroad pillars were
shattered. and subsidence bounds have been described as sharp scarps (23).
A number of mechanisms, including a chain reaction pillar
Surface subsidence of 0.75 to 0.9 m (2.5 to 3 ft), with a collapse have been proposed (21; 22; 24). Most recently, Board et
maximum of 1 m (3.3 ft), was noted over the collapse area (22) al. (23) attribute the failure to “violent shear failure of a thick and
strong overburden bed that was capable of application of full ¾ The Teutschenthal potash mine collapse of Sept. 11, 1996
overburden loading over the entire pillar geometry.” involved failure of 700 long pillars over an area of 2.5 km2
and under 620 to 770 m of overburden in approximately 2
Pechmann et al. (20) successfully fit the seismic event to a seconds. The collapse produced a 4.8 magnitude seismic
crack closure (implosional) mechanism, the crack being the event and 0.5 m of surface subsidence (28).
mined trona seam. Seismic energy released was about 10% of the
potential energy lost by observed subsidence of the overburden An additional event, and the largest of these, is the March 13,
(25). A near-vertical shear failure of the sandstone is consistent 1989 collapse of the Merkers mine, 750-900 m beneath the town
with both brittle failure and a sharp subsidence scarp. In such a of Volkershausen. The collapse involved an area of 6.5 km2,
case, both pillars support and the bridging Tower Sandstone produced a 5.6 magnitude seismic event* and caused
failed, and failed quickly during the collapse. “catastrophic” damage to the town (30). Two levels were mined.
A primary level in carnallite was extensively mined, leaving 30
German Potash Mine Collapses by 6 m pillars with a width to height ratio of 4 to 7. A secondary,
overlying seam was mined in hard rock salt to a lesser extent,
German potash mines are often overlain by sandstone strata initially leaving large, stable pillars. In 1987, mining in the upper
regardless of whether the evaporites are within stratabound or seam was modified to leave yield pillars with a width to height
domal geometries as in Figure 7. Beneath the sandstone, there are ratio of 1.7:1. Heavily “working” rock and roof control problems
typically several hundred meters of “yield, i.e., creepable, were encountered. The actual event appeared to originate with
saliferous rock layers, which hold the minable seam horizons of blasting of carnallite in the lower seam within the pressure
the potash mining” (26). Failure is characterized as involving “the abutment of the upper seam. Failure of 3200 pillars within the
entire structure” with caving up to the surface (subsidence) and lower seam occurred within a time span of 2 to 3 seconds,
“tremors, similar to those experienced during an earthquake.” resulting in up to 1 m of surface subsidence.

The ground subsidence area was described as “rather sharply

bounded by known or inferred fault or shear zones, and pillars on
the other side of these faults remained intact” (12). Damage to
structures straddling the shear zone was extensive (Figure 8).
These faults and shear zones apparently weakened the bridging
Buntsandstein sandstone, about 200 meters above the mining
level. Injection of wastewater into this and adjacent dolomite
strata may have contributed to weakening of these features.

Figure 7. Example of German salt dome potash occurrence

with capping sandstone strata. Salzdetfurth mine (29; 12).

Many collapses have occurred during mining of carnallite

potash seams. Generally, the lower the halite content of a rock
(with carnallite replacing halite) the more brittle its behavior.
Initiation of collapse is generally attributed to a blasting event that
begins a chain-reaction failure incorporating both pillars and roof
strata, resulting in significant surface subsidence and a large
Figure 8. Damage to structure in Volkershausen
seismic event. These collapses, named after overlying villages
(27), include: directly above the shear zone (27).

* Descriptions of seismic event(s) vary. Garrett (12) describes the

¾ A collapse at Teutschenthal in 1940 under 730 m of collapse as occurring in “three major shocks over a 5 second
overburden (26). Failure propagated 1300 m in 0.9 seconds
period of 4.4, 5.1 and 5.5.” Bennet and McLaughling (31)
through slender pillars with a width:height ratio less than
described the seismic event as having a seismic magnitude of 5.4,
one. Back-calculation estimated a local magnitude of 5 for
representing 0.5% of the potential energy lost through subsidence.
the resulting seismic event.
They also found the seismic characteristics of this event to be
¾ A collapse at Herringen with a local seismic magnitude of
quite similar to that produced by the 1995 Solvay event.
5.
¾ Collapses at Merkers on July 8, 1958 and June 29, 1961
Discussion. The value of a set of case studies like these lies in
with seismic magnitudes of 4.8 and 3.7, respectively.
their ability, through a composite view, to more fully reveal the
¾ Collapse at Suenna in 1975 with a seismic magnitude of
mechanisms at work. The most important observation is that there
5.2. This collapse was located in the Werra potash district
is a strong correlation between dynamic collapse events and the
near Suenna, Germany and involved crushing of carnallite
presence of relatively strong, brittle bridging strata above the bed
pillars over an area of 3.35 km2 (12). Damage was entirely
being mined. Other characteristics of these events, including
contained within known fault planes.
collapse mode, seismic energy and the interplay of creep, energy mechanism has also been explored in physical models (Figure
and dynamic failure are explored further. 11).

Collapse Mode Board et al. (23) used the ground reaction curve concept to
explore this mode of strata collapse. They modeled the panel as
Two collapse mechanisms have been evoked in explanations supported by a pressure that was slowly reduced from in situ
of these cases. The first is a cascading or chain reaction failure of stress levels to pressures equivalent to pillar support with
mine pillars. The second is sudden vertical shear failure of strong, increasing extraction ratios. At some point, stable redistribution of
brittle overburden strata. stress to barrier pillars is interrupted by sudden failure of bridging
strata (Figure 12). Board et al argue that panel pillar design must
Whether or not collapse is controlled by chain-reaction failure preserve the integrity of bridging overburden and avoid the
of pillars can be addressed most directly by examining cases unstable ground reaction curve associated with its failure.
where there are no pillars. For evaporite mining, such cases are
conveniently created by solution mining. Daupley et al. (4) Seismic Energy
explored collapse mode for various overburden (Figure 9). They
found that generally soft strata were associated with slow The seismic event created by sudden collapse carries
development of a subsidence trough while the presence of information on the source mechanism of the event. The most
stronger strata was associated with sudden subsidence events. comprehensive seismic analyses in this review were conducted
Failure of strong strata was found to occur on near-vertical shears. for the Solvay case and include two important findings. First, the
first motion of the ground recorded by seismographs is
The vertical shear mechanism was briefly explored in a two- downward, indicating a collapse or implosion often described as a
dimensional numerical model (Figure 10). The numerical model horizontal crack closure motion. Second, the potential energy
used a strain softening constitutive law for the strong, brittle released through subsidence is sufficient to produce the observed
stratum. Softer, weaker overlying strata bend rather than fracture, seismic event.
leaving a more gradual slope. The vertical shear failure

Figure 9. Relationship between strata movements over time and overburden stratigraphy above solution-mined salt seams.
Sudden subsidence caused by caving of stiff layer is often accompanied by a sizable seismic event. After Daupley et al. (4).
 Figure 10. FLAC model showing development of vertical
shears along the face of abutments and tensile failure
in the center of the roof beam.

Figure 13. Seismic magnitude, a logarithmic measure, plotted

against relative change in potential energy of the
subsiding block.

collapse, weakening defects in bridging strata were discovered

during post-failure analyses. The importance of such defects
highlights the critical role of the strata bridge. That is, these
defects violate a fundamental assumption in the design of yield
pillar panels; that overlying strata can bridge the panel, shifting
all but immediate overburden loading to adjacent barrier pillars.

Creep, Energy and Dynamic Failure

The energy approach begins with the fact that underground

mining, by creating the opportunity for overlying strata to fall,
Figure 11. Physical model of vertical shear failure (32). creates substantial amounts of potential energy. This energy may
be contained in perpetuity if sufficient support is provided.
Otherwise, it must be expended either in gradual closure of mined
spaces or in a sudden collapse. The magnitude of available energy
can be immense, as demonstrated by the magnitude of seismic
energy released during collapse events. Management of this
energy and its release is essential to prevention of violent mine
failure.

Highly stressed evaporite minerals typically deform as a

viscoplastic material, gradually dissipating potential energy.
Creep deformation sheds load from highly stressed pillars and
abutments towards other portions of the rock mass. This behavior
is advantageous for yield pillar panels, realizing good initial
support pressure with high extraction ratios, and has been applied
Figure 12. Stable and unstable ground reaction curves for widely in evaporite mining. However, the support capacity of
strata bridging a pillar mining panel. The unstable curve evaporite pillars, particularly yield pillars, will degrade with time,
reflects failure of bridging overburden strata (23). particularly in the presence of groundwater. Since pillars are the
main difference between the work of Daupley et al., (4) on
solution mining and conventional underground mining,
The consistency of potential energy changes and seismic degradation of yield pillars will serve to increase their similarities
energy release was examined. Estimates of the relative change in over time.
potential energy, based on the volume and displacement of
collapsing overburden, were made for several large collapses and Most evaporites also accumulate damage during creep and will
compared to reported seismic energy. The linearity of these eventually lose cohesion and fail. Moreover, creep may shift
results (Figure 13), shows that seismic energy release scales stress to and through structures (faults, strong sandstone beds,
consistently with changes in potential energy loss. This etc.) that are inherently brittle. Differential deformation between
consistency suggests a similar consistency in failure mechanism. the evaporite deposit and overlying strata can also form voids,
which may fill with pressurized gas and/or fluid. Failure of brittle
The crack closure (implosional) character of seismic records elements, particular failures that compromise bridging of
from these events does not provide good insight into whether the overburden stress, can exert a sudden dynamic load on pillars,
event was triggered by failure of a pillar or strata. The short resulting in their sudden failure.
duration and wide extent of these events suggest that strata failure
occurred early, as either the initiator or immediately after the In some cases, creep contributes to the heightened level of
initiating failure, providing a shock load that rapidly drove failure energy release by providing a temporary stability, albeit one that
throughout the collapse area. In some cases, as in the Retsof is inherently unsustainable. Temporary stability provides an
opportunity to extract more resource, increasing the magnitude of The cases reviewed here show a clear association between
the eventual failure. Properly managed, such a process can violent collapse and strong overburden strata. In these cases,
provide high extraction and completion of mining before collapse. failure of bridging overburden creates a shock load that drives
However, projecting when collapse will become imminent is rapid failure of pillars, all within a very few seconds. These cases
problematic and subject to many geologic uncertainties. Even if also show that collapse events can occur where the mined horizon
collapse is delayed until after the mine is abandoned, there may is not brittle and pillars are properly sized. Brittle overburden is
still be consequences for surface structures and ground water. sufficient. Yield pillars in ductile seams may delay collapse,
Improperly managed, the consequences can be substantial, as allowing further mining that increases the potential energy
these case studies have shown. Total extraction in a longwall released in the eventual collapse.
configuration will also induce collapse, but the lack of pillar
support will cause collapse to start earlier and occur in smaller This conclusion concurs with a similar association recently
increments. reported for pillarless solution mines and a recently published
analysis of the Solvay collapse. The conclusion applies to any
The relationship between extraction ratio and the potential mine operating under sedimentary overburden with strong strata.
violence of failure is illustrated conceptually in Figure 14. The
“first cave” line represents the extent of mining before first cave. This conclusion is inconsistent with the cascading pillar
Overburden strata, depth, etc. will define this curve for a given collapse mechanism that has been proposed for some of these
site. Support (pillars, etc.) can extend this span, pushing the mine cases. The distinction between overburden and cascading pillar
along this curve to the right. However, if these supports fail to failure is an important one. The potential for cascading pillar
provide indefinite support, the eventual failure will be more failure is evaluated and controlled by application of pillar design
violent. Subsequent caves extend the cave zone as cantilevering tools. Strata failure, on the other hand, is addressed by panel
strata fail. This caving is represented by horizontal lines design, especially analysis of bridging strata spans and loads. The
extending to the right from the first cave line whose intensity is, design can either protect these spans or, in caving methods, assure
once again, dependent on local overburden, support and stress early and sustained failure. In addition, the overburden failure
conditions. Provision of temporary support with creeping pillars, mechanism identified depends primarily on the nature of
etc. can extend the cantilever distance and increase the violence overburden strata, not the commodity being mined.
of eventual failure of this subsequent caving as well. In other
words, the intensity of caving can be reduced by early and Recognition of the presence of strong strata and proper
sustained caving. consideration of its potential for failure is essential to proper
management of the potential energy reservoir created by mining.
Such recognition is important since many contemporary design
tools, especially those based on empiricism and boundary element
formulations, address only the pillar run mechanism.

ACKNOWLEDGEMENTS

Ron Hughes, General Manager of the Solvay Mine, and Dr.

Mark Board, Itasca Consulting Group are acknowledged for many
helpful discussions in developing the focus of this paper.
Assistance provided by Dr. Mark Larson (NIOSH), who
Figure 14. The intensity of energy release during first contributed the collapse model, is also gratefully acknowledged.
cave increases as a function of span. Early and sustained The views expressed herein are entirely those of the authors and
caving limits the intensity of cave events. do not necessarily reflect those of NIOSH or those who have
provided assistance.

Strong overburden also complicates longwall mining, REFERENCES

interfering with the steady progression of caving. In such cases,
caving may need to be induced artificially or, alternatively, a 1. Duncan, E.J. Scott and Emery Z. Lajtai, The Creep of
bridge of strata maintained over the panel. Permanent Potash Salt Rocks From Saskatchewan. Geotechnical and
maintenance of bridging strata requires a substantial safety factor, Geological Engineering, 1993, 11, 159-184.
purchased by a substantial loss of resource to supporting barrier
pillars and a thorough geomechanical exploration of the 2. Berest, P., B. Brouard and B. Feuga, (2003) Dry mine
overburden bridge. Backfilling and/or brine flooding may abandonment, Solution Mining Research Institute Technical
enhance long-term stability (33). Conference Paper, Fall 2003 Conference, 5-8 October,
Chester, United Kingdom, 28 pp.

CONCLUSIONS 3. Daw, G.P. and C.A. Pollard, Grouting for Ground Water
Control in Underground Mining. International Journal of
Undermining of strata creates a significant reservoir of Mine Water, 1986, vol. 5 (4), Springer, pp. 1-40.
potential energy that may be supported indefinitely, subsided
gradually over time or released in sudden collapse. Maintaining 4. Daupley, X., H. Cuche, and M. Ghoreychi, Typology of
the safety of miners depends on properly anticipating whether a Strata Movement Related to Old Solution Mining of Salt at
sudden collapse is possible – and if so, proper planning of mine Sarralbe (Lorraine, France). Post-Mining 2005, Nov. 16-17,
operations. Nancy, France, 13 pp.
5. Kupfer, D. H., Geology and History of the Belle Isle Salt 19. Andrejchuk, V., Collapse Above the World’s Largest
Mine, the Southernmost Five Islands Salt Dome, St. Mary Potash Mine (Ural, Russia). Int. J. Speleol., 2002, 31 (1/4)
Parish, Louisiana, Louisiana Geological Survey, 1998. 2002: 137-158.

6. Hoffian, E.L. and B.L. Ehgartner, 1996. Three 20. Pechmann, J.C., W.R. Walter, S.J. Nava and W.J.
Dimensional Finite Element Simulations of Room and Pillar Arabasz, The February 3, 1995, ML 5.1 Seismic Event in the
Mines in Rock Salt, SAND96-0988C. June 1996. Presented Trona Mining District of Southwestern Wyoming. Preprint
at Second North American Rock Mechanics Symposium for Seismological Research Letters, Vol. 66, No. 3, May
Montreal, Quebec, June 9-21, 1996. Sandia National 1995, 14 pp.
Laboratories, Albuquerque, NM.
21. Swanson, P.L. and F.M. Boler, (1995) The magnitude 5.3
7. Bauer, S.J., B. L. Ehgartner, and J.T. Neal, Geotechnical seismic event and collapse of the Solvay Trona Mine:
Studies Associated with Decommissioning the Strategic Analysis of pillar/floor failure stability. USBM open file
Petroleum Reserve Facility at Weeks Island, Louisiana : A report 86-95, 82 pp.
Case History. Proc. Technical Class and Technical Session
SMRI Fall Meeting, 2000, San Antonio, p. 146-156. 22. Ferriter, R.L., R.K. Zipf Jr, D.M. Ropchan, J. Davidson,
(1996). Report of Technical Investigation, Underground
8. Hasagawa, H.S., R.J. Wetmiller and D.J. Gendzwill, Nonmetal Mine, Mine Collapse Accident, Solvay Mine,
Induced seismicity in mines in Canada – An overview. Solvay Minerals, Inc., Green River, Sweetwater County,
PAGEOPH, 1989, Vol. 129, Nos. 3/4. pp. 423-453. Wyoming, Gebruary 3, 1995. Denver CO: U.S. Department
of Labor, Mine Safety and Health Administration.
9. Horner, Robert B., Earthquakes in Saskatchewan: A
Potential Hazard to the Potash Industry. Potash ’83., 1983, 23. Board, M, B., Damjanac and M. Pierce (2007) Development
pp. 185-190. of a methodology for analysis of instability in room and
pillar mines. In Deep Mining 07, Proceedings of the Fourth
10. Sepehr, K. and B. Stimpson, Induced Seismicity in Potash International Seminar on Deep and High Stress Mining,
Mining – A Finite Element Study. Int. J. of Mining and Yves Potvin (ed), pp. 273-282.
Geological Engr., 1988, vol. 6, pp. 27-40.
24. Zipf, R.K., Pillar Design to Prevent Collapse of Room-and-
11. Fuzesy, A., Potash in Saskatchewan, Saskatchewan Pillar Mines. Chapter 59 in Underground Mining Methods:
Geological Survey, 1982, Report 181, 44 pp. Engineering Fundamentals and International Case Studies,
2001. W. Hustrulid and R. Bullock (eds), SME, pp. 493-511.
12. Garrett, D.E., Potash: Deposits, Processing, Properties and
Uses. Chapman & Hall, 1996, p. 289. 25. Zipf, R.K. Jr and P. Swanson, Description of a Large
Catastrophic Failure in a Southwestern Wyoming Trona
13. Chen, H. and S. Peng, Mine Panel Collapse – Two Case Mine. Rock Mechanics for Industry, Amadei, Kranz, Scott &
Studies. 18th Conference on Ground Control in Mining, Smeallie (eds), Balkema, 1999, pp. 293-298.
1999, S. Peng (ed.), West Virginia University, pp. 29-39.
26. Minkley, W. and W. Menzel, Dynamic System Stability of
14. Gowan, S.W. and S.M. Trader, Mine Failure Associated Mining Structures in Salt Mining. Rock Mechanics: Tools
With a Pressurized Brine Horizon; Retsof Salt Mine, and Techniques; Proceedings of the 2nd North American
Western New York. Environmental and Engineering Rock Mechanics Symposium: NARMS ’96, Montreal,
Geoscience, 2000; February 2000, v. 6, no. 1, p. 57-70. Canada, 19-21 June 1996. Aubertin, Hassani and Mitri
(eds.). Balkema, pp. 93-100.
15. Kappel, W.M, R.M. Yager and T.S. Miller, The Retsof
Salt Mine Collapse. Land Subsidence in the United States. 27. Minkley, W. and J. Mühlbauer, Re-Stabilization of the
D. Galloway, D. Jones and S. Ingebritsen (eds), USGS Carnallitic Working Areas in the Merkers Potash Mine. Post-
Circular 1182, 1999, pp. 111-120. Mining 2005, November 16-17, Nancy, France.
16. Van Sambeek, L.L., S.W. Gowan and K.A. Payment, 28. Minkley, W., W. Menzel, H. Konietzky and L. te Kamp,
Loss of the Retsof Salt Mine: Engineering Analysis. Proc. of “A Visoco-Elasto-Plastic Softening Model and Its
8th World Salt Symp, Elsevier, 2000, pp. 411-416. Application for solving Static and Dynamic Stability
Problems in Potash Mining,” in FLAC and Numerical
17. Malovichko, A., D. Shulakov, R. Dyaguilev, R. Sabirov Modeling in Geomechanics – 2001 (Proceedings of the 2nd
and B. Ahmetov, Comprehensive Monitoring of the Large International FLAC Conference, Lyon, France, October
Mine-Collapse at the Upper Kama Potash Deposit in 2001), D. Billaux et al., Eds. Lisse: Balkema, pp. 21-27.
Western Ural. Rockbursts and Seismicity in Mines –
RaSiM5, South African Institute of Mining and Metallurgy, 29. Smith, J.P., Notes on the Geology of the Potash Deposits of
2001, pp. 309-312. Germany, France and Spain, Am. Inst. Min. & Met. Engr.,
Transactions, 1951, V. 187, pp. 117-121.
18. Malovichko, A.A., R.H. Sabirov, and B. Sh. Akhmetov,
Ten Years of Seismic Monitoring in Mines of the 30. Minkley, W., (2004) Back analysis rock burst
Verkhnekamskoye Potash Deposit. In Controlling Seismic Völkershausen 1989. In Numerical Modeling of Discrete
Risk: Proceedings of the Sixth International Symposium on Materials – Konietzky (ed.), pp. 105-112.
Rockbursts and Seismicity in Mines (RaSim 6), 9-11 March
2005, Australia, pp. 367-372.
31. Bennett, T.J. and K.L. McLaughlin, Seismic 32. Cloos, H., Zur Experimentellen Tektonik. Geol. Rundschau
Characteristics and Mechanisms of Rockbursts for Use in 21, 1930, pp. 353 – 367.
Seismic Discrimination. In Rockbursts and Seismicity in
Mines: Proceedings of the 4th International Symposium on 33. Pierre, B., B. Benoit and F. Bernard, Dry Mine
Rockbursts and Seismicity in Mines, Krakow, Poland, 11-14 Abandonment. Solution Mining Research Institute, Fall
August, 1997, A.A. Balkema, pp. 61-66. 2003, Technical Meeting, Chester, UK, 28 pp.