Content uploaded by Aleksander J. Mendecki
Author content
All content in this area was uploaded by Aleksander J. Mendecki on Feb 04, 2015
Content may be subject to copyright.
Australian Earthquake Engineering Society 2010 Conference
Perth, Western Australia
Routine Micro-Seismic Monitoring in
Mines
Aleksander J. Mendecki, Richard A. Lynch & Dmitriy A. Malovichko
Institute of Mine Seismology
November 2010
Abstract
Routine seismic monitoring in mines enables the quantification of
exposure to seismicity and provides a logistical tool to guide the effort
into the prevention and control of, and alerts to, potential rock mass
instabilities that could result in rock bursts. One can define the fol-
lowing five specific objectives of monitoring the seismic response of
the rock mass to mining: rescue of personnel, prevention,seismic
hazard rating,alerts - including short term response to unexpected
strong changes in certain parameters - and back analysis to improve
the efficiency of both the mine layout design and the monitoring pro-
cess. A quantitative description of seismic events and of seismicity
are necessary, but not sufficient, in achieving the above objectives.
The paper describes the basis of a modern digital seismic technol-
ogy and seismological parameters used to quantify seismic sources
and seismicity for seismic hazard assessment and rock mass stability
analysis.
1 Objectives of Seismic Monitoring in Mines
In general, routine seismic monitoring enables the quantification of ex-
posure to seismicity and provides a logistical tool to guide the effort into
1
Routine Micro-Seismic Monitoring in Mines 2
prevention, rating of seismic hazards and alerts to potential rock mass in-
stabilities that could result in rock bursts in underground mines (Mendecki,
1997b) and slope instabilities in open pit mines (Lynch et al., 2005; Lynch
and Malovichko, 2006; Malovichko and Lynch, 2006) One can define the
following five specific objectives of monitoring the seismic response of the
rock mass to mining (Mendecki et al., 1999).
1. Rescue: To detect and locate potentially damaging seismic events,
to alert management and to assist in rescue operations.
2. Prevention: To compare the observed and the expected seismic
rock mass response to mining. To confirm the rock mass stability
related assumptions made during the design process and enable an
audit of, and corrections to, the particulars of a given design while
mining.
3. Seismic Hazard Rating: To quantify the exposure to seismicity and
to monitor its spatial and temporal changes. To classify the observed
spatial and temporal seismic patterns into an agreed seismic hazard
rating system.
4. Alerts: To detect strong and unexpected changes in the spatial and/or
temporal behaviour of seismic parameters that could lead to instabil-
ity, affecting working places immediately or in the short term.
5. Back Analysis: To improve both the mine design and the seismic
monitoring processes. Particularly important is thorough and objec-
tive back analysis of larger rock mass instabilities even if they did
not result in injuries, loss of life or damage. Back analysis should
form the basis for a regular critical review of the applied seismic risk
management strategy, guidelines and procedures.
A quantitative description of seismic events and seismicity is considered
the minimum requirement to achieve the above objectives.
2 The Importance of Location
The location of a seismic event is assumed to be a point within the seis-
mic source that triggered the set of seismic sites used to locate it. The
complexity of processes at the source, however, may complicate the loca-
tion of a seismic event. If a slow or weak rupture starts at a certain point,
the closest site(s) may record waves radiated from that very point while
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 3
others may only record waves generated later in the rupture process by a
higher stress drop patch of the same source. One needs to be specific in
determining the arrival times of different phases if the location of rupture
initiation is sought, otherwise the location will be a statistical average of
different parts of the same source.
A reasonably accurate location is important for the following reasons :
•To indicate the location of potential rock bursts.
•All subsequent seismological processing, e.g. quantification of seis-
mic sources, attenuation or velocity inversion, depends on location.
•All subsequent interpretation of individual events depends on loca-
tion, e.g. events far from active mining, close to a shaft or, in general,
in places not predicted by numerical modelling, may raise concerns.
•All subsequent interpretation of seismicity, e.g. clustering and specif-
ically localization around planes, migration, spatio-temporal gradi-
ents of seismic parameters and other patterns are judged by their
location and timing.
Since the source of a seismic event has a finite size, the attainable location
accuracy of all seismic events in a given area should be within the typical
size of an event of that magnitude which defines the sensitivity of the seis-
mic network for that area, i.e. the minimum magnitude, mmin, above which
the system records all events with sufficient signal to noise ratio, see Table
1.
Table 1: Recommended minimum location accuracy for different network
sensitivities as defined by the mmin. Approximate source sizes for a given
mmin for stress drops 0.5 MPa and 0.1 MPa, are also given as a reference.
Sensitivity, mmin , 1.0 0.5 0.0 -0.5 -1.0 -1.5 -2.0
source size, [m] 65-110 35-64 20-35 12-20 6-12 4-6 2-4
Location Accuracy, [m] 100 75 40 20 15 10 5
Given the high quality data from at least 6 three component sites of
reasonable configuration, the error may be reduced to less than 3% of the
average hypocentral distance to the sites used in the location.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 4
3 Quantification of Seismic Sources
A seismic event is considered to be described quantitatively when apart
from its timing, t, and location, X= (x, y, z), at least two independent
parameters pertaining to the seismic source namely, seismic potency, P,
which measures co-seismic inelastic deformation at the source and radi-
ated seismic energy, E, are determined reliably.
Seismic Potency. Seismic potency Prepresents the volume of rock,
of whatever shape, associated with co-seismic inelastic deformation at the
source (Ben-Menahem and Singh, 1981, King, 1978). The scalar seismic
potency [m3], is the product of the strain change and the source volume
P= ∆V. (1)
For a planar shear source, the potency is defined as P= ¯uA, where Ais
the source area and ¯uis the average slip. Pis expressed in [m·m2].
At the source, seismic potency is the integral of the source time func-
tion over the duration. At the recording site potency is proportional to the
integral of the P or S-wave displacement pulse corrected for the far-field
radiation pattern ucorr (t)
PP,S = 4πvP ,S Rˆts
0
ucorr (t)dt, (2)
where vP,S is P or S-wave velocity, Rdistance from the source, tsis source
duration and u(0) = 0 and u(ts)= 0. It is most frequently estimated in
frequency domain from the amplitude of the low frequency displacement
spectra Ω0of the recorded waveforms (Keilis-Borok, 1959)
PP,S = 4πvP ,S RΩ0,P,S
ΛP,S
(3)
where ΛP,S is the root-mean-square value for the radiation pattern of far-
field amplitudes averaged over the focal sphere and ΛP= 0.516 for P-wave
and ΛS= 0.632 for S-wave (Aki and Richards, 2002).
Seismic Energy. The energy release during fracture and frictional slid-
ing is due to the transformation of elastic strain into inelastic strain. This
transformation may occur at different rates ranging from slow creep-like
events to very fast dynamic seismic events with an average velocity of de-
formation at the source of up to a few meters per second. Slow type events
have a long time duration at the source and thus radiate predominantly
lower frequency waves, as opposed to dynamic sources of the same size.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 5
Since excitation of seismic energy can be represented in terms of tempo-
ral derivatives of the source function one may infer that a slower source
process implies less seismic radiation. In terms of fracture mechanics,
the slower the rupture velocity, the less energy is radiated; the quasi-static
rupture would radiate practically no energy.
In time domain the radiated seismic energy of the Por S-wave is pro-
portional to the integral of the radiation pattern corrected far-field velocity
pulse squared ˙u2(t)of duration ts,
EP,S =8
5πρvP,S R2ˆts
0
˙u2
corr (t)dt, (4)
where ρis rock density, vS,P is S or P-wave velocity and Ris the distance
from the source. In the far field of seismic observations the P and S-wave
contribution to the total radiated energy are proportional to the integral
of the square of the P and S velocity spectrum. For a reasonable sig-
nal to noise ratio in the bandwidth of frequencies available on both sides
of the dominant (corner) frequency f0, the determination of that integral
from waveforms recorded by seismic network is fairly objective. The high
frequency component of seismic radiation needs to be recorded by the
seismic system if a meaningful insight into the stress regime at the source
is to be gain.
Apparent Stress. The apparent stress σAis defined as the ratio of the
radiated seismic energy Eto potency P
σA=E
P(5)
and it measures the amount of radiated seismic energy per unit volume of
inelastic deformation at source (Aki, 1966; Wyss and Brune, 1968).
Energy Index. The notion of comparing the radiated energies of seis-
mic events of similar potency can conveniently be translated into a practi-
cal tool called the energy index, EI (van Aswegen and Butler, 1993). The
energy index of an event is the ratio of the observed radiated seismic en-
ergy of that event E, to the average energy ¯
E(P)radiated by events of the
observed potency Ptaken from log ¯
E=dlog P+c, for the area of interest
- Figure 1
EI =E
¯
E(P)=E
10dlog P+c= 10−cE
Pd,(6)
which for d=1.0 would be proportional to the apparent stress. The higher
the energy index the higher the driving stress at the source of the event at
the time of its occurrence.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 6
Figure 1: Energy index concept, after Mendecki (1997a)
Apparent Volume. Source volume can be estimated from, V=P/∆,
and it measures the volume of rock with the inelastic shear strain change
∆. While potency can be reliably derived from waveforms, the strain
change, ∆=c·P f 3
0, is model dependent and it suffers from its cubic de-
pendency on corner frequency f0. Therefore a relatively small uncertainty
in focan cause a large uncertainty ∆and, consequently, in the source
volume V. More stable parameter then is the apparent volume (Mendecki,
1993) that replaces the strain drop with apparent strain:
VA=P
A
=µP 2
E.(7)
where µis the rigidity of the rock. Apparent volume, like apparent stress,
depends on seismic potency and radiated energy, and, because of its
scalar nature, can easily be manipulated in the form of cumulative or con-
tour plots.
Stress and Strain Change. A seismic system can measure only that
portion of strains, stresses or rheology of the process which is associ-
ated with recorded seismic waves. The wider the frequency and amplitude
range and the higher the throughput of the system, the more reliable and
more relevant the measured values of these parameters become.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 7
Seismic waveforms provide information about the strain and stress
changes at the source. The source of a seismic event associated with
a weaker geological feature or with a softer patch in the rock mass yields
more slowly under lower differentiial stress, and radiates less seismic en-
ergy per unit of inelastic co-seismic deformation, than an equivalent source
within strong and highly stressed rock. By comparing radiated seismic en-
ergies of seismic events with similar potencies one can gain insight into the
stresses acting within the part of the rock mass affected by these events -
see Figure 2.
Figure 2: Apparent stresses of selected seismic events with equal local
magnitudes m= 1.0 in a South African Gold Mine. The ratio of the highest
to the lowest apparent stress on this figure is nearly 100.
The routine estimates of seismic potency and radiated seismic energy
from waveforms are relatively inaccurate, with uncertainties, as measured
from the scatter of processed data around the model, from 50% for well be-
haved waveforms to over 100% for complex ones. However, the variation
in radiated seismic energy of seismic events with similar potency occurring
in different stress and/or strain regimes at the same mine is considerably
greater than the uncertainty in measurements and the error propagation
in processing. Thus, while these uncertainties influence the resolution ob-
tainable negatively, they should not prevent the quantitative interpretation
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 8
and comparison of seismic strain and stress changes between different
time intervals and/or between different areas covered by the same seismic
system.
Having recorded and processed a number of seismic events within
a given volume of interest ∆Vover time ∆t, one can then quantify the
changes in the strain and stress regimes and in the rheological properties
of the rock mass deformation associated with the seismic radiation.
This presents an opportunity to validate the results of numerical mod-
elling of the design process. In the numerical modelling practice the as-
sumption of the same elastic constants within a given volume of rock,
σ=constant ·, makes strain and stress distribution equivalent. However,
seismically inferred stress and strain changes are independent. Seismic
strain associated with seismic events in a given volume is proportional to
seismic potencies, s∼PP, and seismic stress is proportional to the ratio
of seismic energies to seismic potencies, σs∼PE/ PP. Since seismic
potency and the radiated energy are independent, the contours of seismic
strain and seismic stress may be qualitatively different, reflecting differ-
ences in stress regime and/or rock mass properties - see Figure 3 as an
example. It is the difference between the modelled stress and/or strain dis-
tributions and the observed ones that need to be explained and reconciled
with the design while mining (Mendecki et al., 1999).
Figure 3: Contours of seismic strain (left) and seismic stress (right) for
seismic events of all magnitudes in the area shown in Figure 1. Please
note the qualitative difference between the distributions of seismic strains
and stresses.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 9
4 Seismic Source Mechanism
4.1 Representation of seismic sources
The low frequency component of seismic waves generated by the majority
of dynamical processes in rock masses can be described using a set of
force dipoles acting at some point within the elastic source (Backus and
Mulcahy, 1976). The exceptions are processes accompanied by move-
ment of mass in the source - this seismic radiation is characterized by a
single force (Takei and Kumazawa, 1994). Thus a set of 9 basic force
dipoles, comprising the seismic moment tensor Mij , is a universal de-
scriptor of seismic radiation for a wide range of source processes includ-
ing mining-induced seismicity. For instance, a planar shear failure may
be described by two dipoles with opposite moments, while radiation of
a symmetric explosive source may be described by three orthogonal lin-
ear dipoles. Another fundamental description of seismic radiation – the
potency tensor Pij - has become widespread of late. This quantity char-
acterises the virtual elastic unconfined deformation of the source mate-
rial which has the same seismic effect as the actual inelastic deformation.
The benefits of Pij over Mij is the fact that the potency tensor doesn’t de-
pend on (essentially unknown) parameters of material in the vicinity of the
source.
Thereby tensors Mij and Pij represent elegant structures providing
two-way connection between source processes and their seismic signals.
On the one hand a set of equivalent force dipoles Mij (or equivalent elastic
deformations measures Pij ) could be specified for practically all processes,
and seismograms could be synthesized for these equivalent dipoles or de-
formations. On the other hand inversion of seismic moment tensor or po-
tency tensor can be performed for recorded seismic waveforms, with the
subsequent benefits of our understanding of the source mechanisms.
4.2 Interpretation of the mechanisms
The initial step of interpretation usually consists of decomposition of tensor
Mij or Pij into two parts. Seismic moment and potency tensors possess
property of symmetry of components for the majority of source processes
(the exceptions can be found again in Takei and Kumazawa, 1994) and as
a result they will have only two parts – the isotropic and deviatoric compo-
nents. The isotropic component characterizes change in volume of mate-
rial in the seismic event source. For instance, in the case of a planar shear
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 10
fault this component will be zero. The deviatoric component describes the
inelastic change of shape of material in the source.
A more specific interpretation of the obtained seismic moment or po-
tency tensor is possible if some specific source model is assumed. For
example, two alternative variants of geometrical parameters of faulting
(orientation of the plane and direction of the faulting) can be estimated
according to the deviatoric components of Mij or Pij if planar shear fault-
ing is assumed to be the source process.
4.3 Visualization and estimation of the mechanisms
The common visualization of seismic source mechanisms is a representa-
tion of P-waves polarities in far-field zone (compressions or dilatations) on
a virtual sphere centered with source (the focal sphere). This sphere with
associated polarities is usually represented in a form of horizontal projec-
tion. In the case of shear faulting the focal sphere is divided by two nodal
planes onto 4 quadrants having alternate polarities (Figure 4a). The re-
gions of compression and dilatation phases will have another configuration
for different types of sources. For example, in the case of complex failure
including simultaneous rupture of shear and tensile cracks the mechanism
will contain isotropic (explosive) component giving rise to larger area of
positive polarities of P-waves in comparison with the area of negative ones
(Figure 4b).
Currently there is a wide spectrum of techniques for the estimation of
seismic source mechanisms for mine seismicity. Some techniques operate
within the bound of specific model of the source – generally this model is a
planar shear faulting (Voinov and Selivonik, 1998; Younga and Fedotova,
2000). Others give the possibility to obtain universal characteristics in the
form of tensors Mij or Pij (Mendecki, 1993; Trifu et al., 2000). Also the
techniques are discriminated according to type of input data – some of
them are based only on polarities and amplitudes of P- and S-waves while
others involve full waveforms.
Some features of mine induced seismicity have aided the estimation of
seismic source mechanisms. Firstly the seismic arrays installed in mines
are usually 3-dimensional, much better than the 2-dimensional arrays used
in crustal and regional seismic monitoring. Secondly the medium may usu-
ally be assumed to be homogeneous and isotropic, unlike the large differ-
ences present over hundreds or thousands of kilometers in the Earth’s
crust.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 11
Figure 4: Visualization of mechanisms of seismic events: the distribution of
first motion polarities on the focal sphere for different source mechanisms
is shown in (a) (from USGS, 1996), while in (b) the source mechanism
representation for the case of combined shear and tensile fractures are
shown (from Julian et al., 1998).
An example of seismic source mechanism invertion for mines is pre-
sented in Figure 5. The mechanisms of 7 large events which took place
during a 6 year period and their 32 aftershocks in a section of a block-
caving mine are shown as “beach balls”. The cave surface is shown in
yellow.
Attention should be paid to two aspects in this example:
•The insignificance of isotropic components in mechanisms of both
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 12
main shocks and aftershocks (the “beach-balls” shown in Figure 5
are similar to ones presented in Figure 4a). This implies that shear
failure dominates the source mechanism.
•The uniformity of the mechanisms. Evidently that the centers of pos-
itive P-waves polarity regions have vertical orientation. The centers
of the negative polarities regions have North-South orientation. Such
uniformity indicates the consistency of factors that initiated these
large events during the long period of time (six years!).
Figure 5: Mechanisms of 7 large seismic events and their 32 aftershocks
in a large block-cave mine, recorded during a 6-year period.
4.4 Application of information about seismic source mech-
anism in mines
The source mechanism and their simplified analysis as demonstrated above
yields new knowledge about seismicity which is complementary to infor-
mation presented in Chapter 3. In general the estimation of seismic source
mechanisms for mine seismicity allows some important practical issues to
be investigated:
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 13
•Investigation of the causes of a specific rockburst. In this case the
parameters of the mechanism of seismic event associated with the
rockburst (orientation of possible nodal planes and slip vectors) are
compared with characteristics of known geological features in the
vicinity of the source.
•Estimation of stress state of rock mass. A collection of geometrical
parameters of the sources of seismic events that occurred in some
rock volume contains information about stress state of this volume,
i.e. about direction of principal axes of stress and ratio of their ampli-
tudes (Dubinski and Stec, 2001).
5 Quantification of Seismicity (Mendecki, 1997a)
The few most frequent quantities derived from waveforms are: time t, lo-
cation X= (x, y, z), potency Pincluding its tensor Pij and radiated energy
E. If seismicity is considered as a group of seismic events confined to a
given volume ∆Vand time ∆tthen, in addition to statistical moments, one
can define a number of parameters based on the following quantities:
•average time between consecutive events ¯
t,
•average distance between consecutive events ¯
X, including source
sizes,
•sum of seismic potencies PPor PPij , or moments
•sum of radiated energies PE.
The following parameters describe the statistical properties of co-seismic
deformation and associated changes in the strain rate, stress and the rhe-
ology of the process. The usefulness of these parameters depends on the
definition of the group of events. It is important that the selection of ∆V
and ∆tbe guided by the characteristic space and time scales associated
with the processes under study.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 14
5.1 Seismic Strain, Stress and Stiffness
The average seismic strain tensor ij , produced by a number of events that
have occurred within the volume ∆Vover the period of time ∆t=t2−t1
is proportional to the sum of their potency tensors (Brune, 1968; Kostrov,
1974; Kostrov and Das, 1988)
sij (∆V, ∆t) = 1
2∆V
t2
X
t1
Pij .(8)
The average seismic strain rate tensor then is
˙sij (∆V, ∆t) = 1
2∆V∆t
t2
X
t1
Pij (9)
The average seismic stress tensor, σsij , has been defined by Kostrov,
1974; Kostrov and Das (1988) as
σsij (∆V∆t) = 1
˙sij ∆V∆t
t2
X
t1
E=2Pt2
t1E
Pt2
t1Pij
.(10)
As tensors, seismic strain, strain rate and seismic stress have the same
principal axes. The ratio of seismic stress to seismic strain can then be
taken as the seismic stiffness modulus Ks,
Ks(∆V, ∆t) = σsij
sij
=4∆VPt2
t1E
Pt2
t1Pij 2.(11)
Seismic stiffness is a scalar and measures the rock mass ability to resist
seismic deformation with increasing stress.
5.2 Seismic Viscosity and Relaxation Time
The rock mass resistance to seismic deformation can also be measured
by seismic viscosity, ηs, defined as the ratio of seismic stress to seismic
strain rate (Kostrov and Das (1988))
ηs(∆V, ∆t) = σsij
˙sij
=4∆V∆tPt2
t1E
Pt2
t1Pij 2νs=ηs
ρ,(12)
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 15
where νsis the kinematic viscosity and ρis density. Elastic properties
govern processes with a characteristic time scale less than the relaxation
time τ=η/µ, [s], where ηis viscosity and µrigidity.
τs(∆V, ∆t) = ηs
µ=4∆V∆tPt2
t1E
µPt2
t1Pij 2(13)
5.3 Seismic Diffusivity and Schmidt Number
The statistical difusivity can be defined as
ds(∆V, ∆t) = ¯
X2
¯
t(14)
where ¯
Xis the mean distance between consecutive events and ¯
tis the
mean time between events. Seismic Schmidt number, that measures the
complexity of seismic deformation, is defined by
Scs(∆V , ∆t) = νs
ds
=4∆V∆t(¯
t)Pt2
t1E
ρ¯
X2Pt2
t1Pij 2(15)
Note that the Schmidt number contains all four basic parameters that
describe seismicity, namely: ¯
t,¯
X,PPand PE.
5.4 Seismicity and Stability
In general the stability of the rock mass subjected to mining can be related
to its stiffness, i.e. its ability to resist deformation with increasing stress.
While the overall stiffness of the rock mass is being maintained the seismic
potency production, PP, is expected to be proportional to the volume
mined, PP∼Vm.
As mining progresses the overall stiffness of the rock mass is being
degraded and the rate of potency production per unit of Vmmay increase.
With further degradation in stiffness, the response may become nonlinear
with accelerating potency release, PP∼(PVmeff )γ, associated with an
increase in activity rate, 1/¯
t, signifying potential for larger instabilities.
The dynamics of such instability expressed by the apparent stress,
σA=E/P , depends on the ratio of the stiffness of the potentially unstable
volume of rock to the stiffness of the surrounding rock mass. The higher
this ratio the more energy will be released per unit of inelastic deformation
at the source.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 16
Given these observations, searches for signs of the stress softening
phase that precedes seismic instability reduces to analysis of the time
histories of energy index and cumulative apparent volume, as proxies for
seismic stress and strain, respectively. A period of increasing energy in-
dex with a normal rate of cumulative apparent volume (the stress hard-
ening phase) followed by a period of dropping energy index and simul-
taneously accelerating cumulative apparent volume (the stress softening
phase) constitutes an indication of potential future instability. With high
resolution micro-seismic arrays, such patterns have been detected before
many large seismic events - see Figures 6 and 7.
Figure 6: The characteristic pattern of dropping energy index and acceler-
ating cumulative apparent volume 30 hours prior to a large seismic event
(local magnitude 2.4 in this case), from seismic data recorded at TauTona
gold mine in South Africa. Data from the period of stress softening (shaded
zone) can be used to spatially identify the zone of the future instability to
within about 100 m - from Lynch and Mendecki (2001).
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 17
Figure 7: The characteristic pattern of dropping energy index and accel-
erating cumulative apparent volume prior to a large seismic event (local
magnitude 1.9 in this case), from seismic data recorded at TauTona gold
mine in South Africa. This time the stress softening phase is detected 6
hours prior to the large event and continues unabated until the instability -
from Lynch and Mendecki (2004).
Another guiding idea in the interpretation of seismic activity is the con-
cept of self-organization into critical state, i.e. a state at which the corre-
lation length becomes comparable with the system size and at which the
system can develop and maintain reproducible relationships among its dis-
tant parts. One would expect an increase in the mean distance between
consecutive events, ¯
X. It is assumed that the growth of long-range cor-
relations within the rock mass allow for progressively larger events to be
generated. Intermittently, correlations may reach or even exceed the size
of the observed area, creating conditions conducive for large instabilities.
At this stage the system’s sensitivity to external or internal disturbances
would diverge offering predictability. The following equations depict the
expected qualitative changes in seismic parameters associated with un-
stable rock mass behaviour.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 18
energy index
apparent volume =&
%=⇓(16)
viscosity :ηs=stress
strain rate =&
%=⇓(17)
diffusivity :ds=(distance between events)2
time between events =%
&=⇑(18)
The Figure 8 depicts an example of the Schmidt number and apparent
volume few hours after the production blasts that resulted in seismic event
of magnitude m= 1.9.
Figure 8: Schmidt number and the cumulative apparent volume over the
24 hours period showing the rock mass response to production blasts at
16h55. The large seismic event m= 1.9 occurred a few hours after the
blasts and was preceeded by drop in Scsand accelerating VA(Mendecki,
2001).
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 19
6 Seismic Monitoring Technology
To achieve meaningful results from analysis of seismic events, the seismo-
gram data itself needs to be accurately recorded. For routine mine seismic
monitoring, typical requirements are for seismic events in the moment-
magnitude1range −2.0≤mHK ≤+3.5to be consistently recorded. Using
the relation derived by Gutenberg and Richter (1956): log E=3
2m+ 4.8,
Table 2 has been computed for reference.
Table 2: Equivalent values of potency, moment-magnitude and radiated
seismic energy using the relations of Hanks and Kanamori (1979) and
Gutenberg and Richter (1956) and assuming a ridigity µ= 30 GPa.
Potency [m3] moment-magnitude Energy [J]
0.000041 -2.0 6.3×101
0.00023 -1.5 3.6×102
0.0013 -1.0 2.0×103
0.0073 -0.5 1.1×104
0.041 0.0 6.3×104
0.23 0.5 3.6×105
1.3 1.0 2.0×106
7.3 1.5 1.1×107
41 2.0 6.3×107
230 2.5 3.6×108
1300 3.0 2.0×109
7300 3.5 1.1×1010
To see what frequency bandwidth would be appropriate to monitor
such typical seismic events, relations from Keilis-Borok (1959) and Brune
(1971) can be combined (Mountfort and Mendecki, 1997) to obtain an ex-
pression for corner frequency, f0, in terms of potency P, stress drop ∆σ,
rigidity µand S-wave velocity vS:
f0=2.34vS
2π
3
s16∆σ
7µP (19)
This relation is presented graphically in Figure 9 using values of S-
wave velocity commonly encountered in underground and open pit mines.
1The moment magnitude or Hanks-Kanamori magnitude scales with seismic potency,
P, and rigidity, µ, as: mH K =2
3log P+2
3log µ−6.06 (Kanamori and Anderson, 1975;
Kanamori, 1977; Hanks and Kanamori, 1979)
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 20
If a typical stress drop of 1 MPa is used, it can be seen that magnitude
-2.0 will produce a corner frequency of approximately 1200 Hz, while a
magnitude +3.5 will produce a corner frequency of approximately 2 Hz.
To accurately estimate potency from the spectral plateau, the seismogram
must be recorded with a frequency content with lower limit at most f0/2.
To accurately estimate radiated seismic energy, the seismogram must be
recorded with a frequency content with upper limit at least 5f0. Thus the
frequency band of the seismic monitoring system (sensors and data ac-
quisition electronics) should be at least 1 Hz to 2400 Hz for common mine
situations.
Figure 9: The relation between S-wave corner frequency and moment
magnitude for a range of stress drops (producing the different series of
lines) and S-wave velocities (producing the thickness of the lines) as com-
monly encountered in mines (Mountfort and Mendecki, 1997).
To estimate what dynamic range is required to record the typical am-
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 21
plitudes of seismic waves, an expression of McGarr (1991) can be used
with equation 19 to produce:
Rvmax =0.0686
ρvS
3
p∆σ2µP (20)
Again using a stress drop of 1 MPa, and typical values of rock density,
S-wave velocity and rigidity, we can see that a magnitude +3.5 event will
produce a peak seismic wave velocity of about 0.1 m/s at a seismic sensor
50 m away. A small event of magnitude -2.0 is estimated to produce a
peak seismic wave velocity of about 10−5m/s at a sensor 500 m from the
source. In practice, this value is observed to be a little lower, at about 10−6
m/s. Since a signal-to-noise ratio of at least 10 is required for accurate
processing of the seismogram data, the seismic system should be able to
reliably record seismic signals of maximum amplitude at least 0.1 m/s, with
a peak noise level of at most 10−7m/s. This implies a dynamic range of the
data acquisition electronics of at least 6 orders of magnitude, which is 120
dB.
Some typical seismograms from micro-seismicity recorded in mines by
the ISS system are shown in Figures 10 and 11. The large range of ampli-
tudes that must be recorded for these seismic events is evident: the small
magnitude -2.0 event generates peak ground motions of about 10−6m/s
at 170 m from the source, while the magnitude +1.8 event generates 10−2
m/s at approximately the same sensor-source distance.
For small events, the seismic signals are both high frequency and low
in amplitude. Surface mounted sensors tend to be noisy, are shielded
from high frequencies by the fracture zone surrounding underground ex-
cavations and suffer from the variable surface amplification effect. Conse-
quently borehole seismic sensors are preferred. The sensors are usually
grouted into holes at least 7 m in depth and of diameter at least 76 mm.
Figure 12 shows an IMS tri-axial 4.5 Hz borehole geophone. Such geo-
phones have a linear response from 4 Hz to at least 2500 Hz. Passive
electronics are fitted to the sensor after installation to provide a frequency-
dependent amplification. This results in an effective lower frequency re-
sponse of about 3 Hz, and does not introduce any noise or distortion of
the signals at higher frequencies.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 22
Figure 10: Seismograms recorded by an array of 14 Hz and 30 Hz bore-
hole geophones for a magnitude -2.0 event.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 23
Figure 11: Seismograms recorded by an array of 14 Hz and 30 Hz bore-
hole geophones for a magnitude +1.8 event.
Figure 12: An IMS tri-axial 4.5 Hz borehole geophone. The sensor is
usually installed in 10 m holes of diameter at least 76 mm. The signals
from such sensors are less noisy and have less distortion than signals
from surface-mounted seismic sensors.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 24
The analogue seismic signals from geophones or accelerometers re-
quire digitisation before the data may be collected and processed on a
computer. In early seismic systems the digitisation was done at a central
point, and the seismic signals were carried from the sensors to this point
in analogue form over copper cables. Unfortunately electrical noise pre-
vents a large ratio between the strongest signals and the noise level, and
so either small signals get obliterated by the induced electrical noise, or
else a pre-amplifier is used in which case the strongest signals become
clipped. Thus these ’analogue’ seismic systems do not have sufficient dy-
namic range to meet the criterion dicussed previously in this section. For
this reason, the ISS system uses distributed seismic stations near each
of the sensors to digitise the signals. Then digital signals are transmit-
ted from the stations to the central point, and so the problem of induced
electrical noise is avoided.
To minimise the requirements of the digital data transmission rate, only
seismograms that have been recorded consistently by a number of seismic
stations are transmitted to the central site. In this way, the sampling rate at
the seismic stations can be much higher than the digital communications
rate to the central site.
The latest GS technology (see Figure 13) is capable of sampling rates
up to 48000 Hz and uses a 32-bit over-sampled analogue-to-digital con-
verter to achieve a dynamic range of 129 dB at 6000 Hz sampling. A digital
sigma-delta anti-aliasing filter provides a useful bandwidth of up to 2700
Hz at 6000 Hz sampling, and so this is sufficient for most routine mine
seismic monitoring purposes.
The power consumption of the GS seismic station is very low at 3 W,
making the unit suitable for battery-powered monitoring applications. An
integrated Uninterruptable Power Supply powers the GS for at least 6
hours in case of power failure, and the seismograms can be stored lo-
cally on a USB flash drive and internal non-volatile memory before the
power is lost. In this way valuable seismograms of a large seismic event
are preserved even if power and communications are disrupted after the
event.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 25
Figure 13: The GS seismic data acquisition unit from ISS International.
With a frequency bandwidth of 0.3 - 48000 Hz and a dynamic range of
129 dB at 6000 Hz sampling rate, this seismic station meets the require-
ments of the typical seismic monitoring system. The unit is compact (140
mm ×200 mm ×65 mm) and has a power consumption of 3 W. Digital
seismogram data is transmitted to the central site and can also be stored
locally on a USB disk.
Most mines around the world do not require seismic-monitoring equip-
ment that is certified to operate in explosive gas (methane) environments.
The GS stations use 3 W of power, and since this much power is suffi-
cient to make an electrical spark, heavy explosive-proof enclosures must
be used for the GS in such environments.
Alternatively, the GSi unit can be used in such mines with explosive
gases present. This unit is a full seismic station, sampling at 5000 Hz with
a 24-bit digitiser and using less than 0.11 W of power. Since the power
consumption is so low, this unit does not require any explosion-proof box
and has been certified for use in Russian and South African mines.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 26
Figure 14: The GSi seismic data acquisition unit from ISS International.
This unit is a full seismic station that uses less than 0.11 W of power and
has been certified for use in explosive gas environments. The unit is very
compact (130 mm ×80 mm ×30 mm) and records digital seismograms
from a tri-axial geophone at a sampling rate of 5000 Hz.
The ISS seismic system is usually permanently installed at locations
underground, as close as practically possible to the working areas given
the constraints of cabling for power and communications. To get seismic
sensors even closer to the zone of rock which is being most influenced
by mining, the GSx wireless seismic stations can be used. These units
have an internal rechargeable battery, digital radio and 14 Hz geophone,
and communicate with the closest GS seismic station using any other
GSx units as relays. This Internet-style of self-configuring communica-
tions makes the units very easy to install, and the internal battery powers
the device for over 3 weeks. The device can measure non-seismic sen-
sors (for example convergence meters) or the uni-axial geophone. For the
seismic data, the signal is sampled and analysed within the GSx unit, and
the micro-seismic activity rate is transmitted back to the central site each
minute.
Digital seismogram data at the central site computer is processed inter-
actively with the JMTS software package to estimate the location, inelas-
tic co-seismic deformation, P, and radiated seismic energy, E, for each
seismic event. Figures 10 and 11 show screen-shots of this software. In
addition, some secondary parameters are extracted for each event, such
as the corner frequency, f0, and inelastic attenuation parameter, Q.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 27
Figure 15: The wireless GSx seismic data acquisition unit from ISS In-
ternational. This unit monitors an internal 14 Hz uni-axial geophone and
reports the micro-seismic activity rate to the closest GS each minute. The
device is powered from a rechargeable internal battery and communicates
via a spread-spectrum digital radio, using other GSx units as relays if nec-
essary. The device has bright lights that can be turned on remotely to warn
local users of unusual micro-seismic activity rates.
These seismological parameters for each event are then analysed, as
discussed in Section 3. Once a group of seismic events have been pro-
cessed and quantified, an analysis of seismicity is possible. This was
described in Section 5 and is achieved using the custom-built 3-D visuali-
sation and seismological analysis software - JDi.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 28
7 Conclusions
This paper has presented the methodology of routine seismic monitoring
for mines, as currently in use at over 150 mines in 30 countries. The
most common objectives of rescue, prevention, seismic hazard rating and
alerting, warning and back-analysis were explored. While monitoring re-
quirements for the objective of rescue is not stringent as only the largest
events need to be located, even very small events must be recorded and
quantified for the objective of warning to be feasible. Thus the required
seismic system must be specified with the overall monitoring objectives in
mind.
In order to achieve any of these objectives, seismic events must be
routinely located in relation to the 3-dimensional geometry of the mine
workings and the source parameters must be quantified. The source size
limits the accuracy of location, as the rupture can start anywhere on the
eventual fault surface. Notwithstanding this, typical 3-D location errors of
3% of the average hypocentral distance should be possible with a modern
seismic monitoring system and well-designed seismic array.
For each event, the radiated seismic energy Eand co-seismic inelastic
deformation Pare independently estimated. From these measurements, it
is possible to quantify the apparent stress drop, energy index and apparent
volume of each seismic event. While the stress drop is commonly consid-
ered, it is however more useful to use energy index as a proxy for the
dynamic stress levels within the rock surrounding a mine. The apparent
volume provides a parameter with information about seismic deformation
that scales more usefully than simple cumulative potency.
Moment tensor estimation is a powerful tool that is often used in seis-
mological back analyses. The understanding of how large seismic events
occur aids in the design of new mine levels in order to minimise the chances
of recurring large seismic events. While this technique has proved to be
the most difficult to implement routinely, new tools are constantly being
developed to make this task simpler.
A group of seismic events can yield interesting conclusions when anal-
ysed jointly. Such is the basis for time history analysis of seismological
parameters like energy index, cumulative apparent volume and seismic
Schmidt number. Stability analyses of mining sections is routinely under-
taken in mines around the world using such time history analysis.
Recently, high precision (10−6) measurements of P-wave velocity vari-
ations have yielded promising precursory patterns prior to some large
(ml= 1.0and ml= 3.0) seismic events on the San Andreas fault (Niu
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 29
et al., 2008). This technique is being attempted in underground mines
now (Lynch, 2010) and, if successful, would further increase the success
rates of advance indications of potentially large instabilities in underground
mines.
Modern digital seismic monitoring technology has evolved over the past
20 years. The large variation in both frequencies and amplitudes of seis-
mic signals from mining induced seismicity means that a large dynamic
range with low noise levels and high sampling rates are required. This
is achieved by the use of distributed seismic stations underground. How-
ever in the mining environment, reliability of underground electronic equip-
ment is of utmost importance. Given the extensive and often quite old
telecommunications network underground, digital modems between dis-
tributed seismic stations and the surface central site must be robust and
fault-tolerant, as well as flexible. These technical challenges drive the de-
velopment of the latest seismic monitoring systems.
As developed as the methodology and technology of mine seismic
monitoring is, the complexity of the subject requires a commitment from
the mine in the form of a dedicated engineer or scientist who will operate
the system. The full benefits of seismic monitoring in mines is attained
when this person is prepared to say “I love my seismic system”(Mikula,
2005).
References
Aki, K. (1966), Generation and propagation of G waves from the Niigata
earthquake of June 16, 1964. Part 2: Estimation of earthquake moment,
released energy, and stress strain drop from the G-wave spectrum, Bul-
letin Earthquake Research Institute Tokyo University,44, 73–88.
Aki, K., and P. G. Richards (2002), Quantitative Seismology, 2nd ed., Uni-
versity Science Books, California.
Backus, G. E., and M. Mulcahy (1976), Moment tensor and other phe-
nomenological descriptions of seismic sources. 1. Continuous displace-
ments, Geophys. J. R. Astron. Soc.,46, 301–329.
Ben-Menahem, A., and S. J. Singh (1981), Seismic Waves and Sources,
Springer-Verlag, New York.
Brune, J. N. (1968), Seismic moment, seismicity and rate of slip along
major fault zones, Journal of Geophysical Research,73, 777–784.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 30
Brune, J. N. (1971), Tectonic stress and the spectra of seismic shear
waves from earthquakes: Correction, Journal of Geophysical Research,
76, 5002.
Dubinski, J., and K. Stec (2001), Relationship between focal mechanism
parameters of mine tremors and local strata tectonics, in Proceedings of
the 5th international symposium on rockbursts and seismicity in mines,
edited by G. van Aswegen, R. J. Durrheim, and W. D. Ortlepp, pp. 113–
118, South African Institute of Mining and Metallurgy.
Gutenberg, B., and C. F. Richter (1956), Magnitude and energy of earth-
quakes, Ann. Geofis.,9, 1–15.
Hanks, T. C., and H. Kanamori (1979), A moment magnitude scale, Journal
of Geophysical Research,84, 2348–2350.
Julian, B. R., A. D. Miller, and G. R. Foulger (1998), Non-double-couple
earthquakes 1. Theory, Reviews of Geophysics,36(4), 525–549.
Kanamori, H. (1977), The energy release in great earthquakes, Journal of
Geophysical Research,82(20), 2981–2987.
Kanamori, H., and D. L. Anderson (1975), Theoretical basis of some em-
pirical relations in seismology, Bulletin of the Seismological Society of
America,65(5), 1073–1095.
Keilis-Borok, V. I. (1959), On estimation of the displacement in an earth-
quake source and of source dimensions, Annali di Geofisica,12(2),
205–214.
King, G. C. P. (1978), Geological faults, fractures, creep and strain, Philo-
sophical Transactions of the Royal Society of London. Series A, Mathe-
matical and Physical Sciences,288(1350), 197–212.
Kostrov, B. V. (1974), Seismic moment and energy of earthquakes and
seismic flow of rock., Izv. Phys. Solid Earth,13, 13–21.
Kostrov, B. V., and S. Das (1988), Principles of Earthquake Source Me-
chanics, Cambridge University Press.
Lynch, R., and D. Malovichko (2006), Seismology and slope stability in
open pit mines, in International Symposium on Stability of Rock Slopes,
International Symposium on Stability of Rock Slopes.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 31
Lynch, R., and A. Mendecki (2004), Gap601a: Experimental and theoreti-
cal investigations of fundamental processes in mining induced fracturing
and rock instability close to excavations - research project for safety in
mines research advisory council, Tech. rep., ISS International Limited.
Lynch, R., and A. J. Mendecki (2001), High-resolution seismic monitoring
in mines, in Proceedings of the 5th International Symposium on Rock-
bursts and Seismicity in Mines, Johannesburg, South Africa, edited by
G. van Aswegen, R. J. Durrheim, and W. D. Ortlepp, pp. 19–24, South
African Institute of Mining and Metallurgy.
Lynch, R., R.A..and Wuite, B. Smith, and A. Cichowicz (2005), Micro-
seismic monitoring of open pit slopes, in Rockbursts and Seismicity in
MInes 6, Perth, edited by Y. Potvin and M. Hudyma, Australian Centre
for Geomechanics.
Lynch, R. A. (2010), Keynote address: Passive and active microseismic
monitoring in mines, in Proceedings of the 5th Deep and High Stress
Mining Seminar, Santiago, Chile.
Malovichko, D. A., and R. A. Lynch (2006), Micro-seismic monitoring of
open-pit slopes, Mining Echo,2(24), 21–30.
McGarr, A. (1991), Observations constraining near-source ground motion
estimated from locally recorded seismograms, Journal of Geophysical
Research,96(B10), 16,495–16,508.
Mendecki, A. J. (1993), Real time quantitative seismology in mines:
Keynote address, in Proceedings of the 3rd International Symposium on
Rockbursts and Seismicity in Mines, Kingston, Ontario, Canada, edited
by R. P. Young, pp. 287–295, Balkema, Rotterdam.
Mendecki, A. J. (1997a), Quantitative seismology and rock mass stability,
in Seismic Monitoring in Mines, edited by A. J. Mendecki, 1 ed., chap. 10,
pp. 178–219, Chapman and Hall, London.
Mendecki, A. J. (Ed.) (1997b), Seismic Monitoring in Mines, 1 ed., 262 pp.,
Chapman & Hall.
Mendecki, A. J. (2001), Data-driven understanding of seismic rock mass
response to mining: Keynote Address, in Proceedings of the 5th Inter-
national Symposium on Rockbursts and Seismicity in Mines, Johannes-
burg, South Africa, edited by G. van Aswegen, R. J. Durrheim, and W. D.
Ortlepp, pp. 1–9, South African Institute of Mining and Metallurgy.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 32
Mendecki, A. J., G. van Aswegwn, and P. Mountfort (1999), A guide to rou-
tine seismic monitoring in mines, in A Handbook on Rock Engineering
Practice for Tabular Hard Rock Mines, edited by A. J. Jager and J. A.
Ryder, chap. 9, pp. 287–309, The Safety in Mines Research Advisory
Committee, Johannesburg.
Mikula, P. A. (2005), The practice of seismic management in mines – how
to love your seismic monitoring system, in Rockbursts and seismicity
in mines: Proceedings of the 6th International Symposium, edited by
Y. Potvin and M. Hudyma, Australian Centre for Geomechanics.
Mountfort, P., and A. J. Mendecki (1997), Seismic transducers, in Seismic
Monitoring in Mines, edited by A. J. Mendecki, 1 ed., chap. 1, pp. 1–20,
Chapman and Hall, London.
Niu, F., P. G. Silver, T. M. Daley, X. Cheng, and E. L. Majer (2008), Pre-
seismic velocity changes observed from active source monitoring at the
parkfield safod drill site, Nature,454.
Takei, Y., and M. Kumazawa (1994), Why have the single force and torque
been excluded from seismic source models?, Geophys. J. Int.,118, 20–
30.
Trifu, C. I., D. Angus, and V. Shumila (2000), A fast evaluation of the seis-
mic moment tensor for induced seismicity, Bulletin of the Seismological
Society of America,90, 1521–1527.
USGS (1996), http://quake.wr.usgs.gov/recenteqs/beachball.html, Tech.
rep.
van Aswegen, G., and A. G. Butler (1993), Applications of quantitative
seismology in South African gold mines, in Proceedings of the 3rd Inter-
national Symposium on Rockbursts and Seismicity in Mines, Kingston,
Ontario, Canada, edited by R. P. Young, pp. 261–266, Balkema, Rotter-
dam, iSBN 90 5410320 5.
Voinov, K. A., and V. G. Selivonik (1998), Estimation of stresses and mobil-
ity of structural blocks according to processing of mechanisms of seismic
event sources, in Mining Geophysics, pp. 86–91, VNIMI.
Wyss, M., and J. N. Brune (1968), Seismic moment, stress and source
dimensions for earthquakes in the California-Nevada region, Journal of
Geophysical Research,73, 4681–4694.
AEES 2010, Perth November 14, 2010
Routine Micro-Seismic Monitoring in Mines 33
Younga, S. L., and Y. V. Fedotova (2000), Study of source mechanisms
of low energy seismic events in khibiny massif, in Proceedings of the
All-Russian Geodynamics and Technogenesis meeting, pp. 171–174.
AEES 2010, Perth November 14, 2010
