The magnitude scales were proposed as an attempt to quantify an earthquake in an instrumental basis. A microseismic is defined as a seismic event with scale magnitude, related to seismic moment, between -2 and 0 (BOHNHOFF et al., 2010). Thus, a monitoring focus on microseismic events must use instruments capable of registering that sensitivity range.

In consequence, the main difference between microseismic and seismic monitoring system are the employed instruments, which are geophones designed to record a range of frequencies adequate to the expected ground motion amplitudes. For microseismic monitoring, the minimum sensitivity is 10-7 mm/s and the sensor’s frequency range vary depending on the natural frequency of the monitored structure. There are mainly two types of sensors:

  • 4.5 ± 0,5 Hz, which registers vibrations from 2.6 Hz to 2000 Hz,
  • 14 ± 0.5 Hz, which registers vibrations from 6 Hz to 2000 Hz.

The geophones are designed in two models – surface and borehole (Figure 1). The surface geophones should be grouted on the rock for a good coupling and the grout should have similar acoustic impedance as the medium (Mendecki, 1999). The borehole geophone should be filled with grout or soil, depending on the installation medium, and the maximum recommended installation depth is 400m (Institute of Mine Seismology, 2021). All sensor must be orientated North. These precautions and complexity are necessary, once, due to their high sensitivity and low theoretical error, uncertainty associated with installation can overcome the method’s uncertainty (Rodrigues, Oliveira, 2020).

The geophones have a smart sensor and an analogic sensor. The smart sensor is responsible for communicating the sensor’s serial identification number and its inclination to the server. The analogical sensor registers the ground motions induced in the medium, in the z- axis (uni-axial geophone) or x- y- z- axis (tri-axial geophone). After the acquisition, the signals are digitalized by an analog to digital convertor. Then, the data is pre-processed by a seismic processor, compressed, synchronized and transferred to the seismic server, to be displayed and processed by a team of trained seismologists in a processing software.

Monitoring approaches

An operational microseismic system works with two different monitoring approaches. The first one, the conventional microseismic, also known as passive microseismic, monitors ground motions with high sensitivity (as low as 10-7 mm/s) induced by seismic sources, locate and estimate source parameters. The second approach uses the seismic ambient noise, the Earth seismic background, as a source. The interferometry is based on the principle of generating a new seismic response for a virtual seismic source and cross-correlating seismic observations at different receiver locations. This process is performed for all pairs of sensors and estimates the wave propagation velocity change between the pairs.

Hypothetical scenario

Considering a earth-filled dam and its reservoir as an area of interest, it would be possible to identify numerous signals, such as a natural source signal (reactivation of fault zone), induced source signal (related to engineering activities – e.g. increase of vertical stress or decrease of effective stress due to increase of pore pressure by reservoir impoundment), mechanical signals (generated by the operation of equipment), electrical signals (spurious signal induced by thunder) and, in the mining industry, blasting signal.

In a scenario in which the monitoring program is focused on induced seismicity by the reservoir and on the response of the structure after the reservoir impoundment, some considerations must be made. As indicated by Barros et al. (2016), an identification of an induced event must be confirmed after positive response to some questions:

  1. Is this the first known event?
  2. Did the events start only after the reservoir was filled?
  3. Is there a clear correlation between the reservoir and the seismicity?
  4. Are the epicenters within the reservoir or a short distance from its surroundings?
  5. Are the epicenters shallow and spatially related to the region affected by the reservoir?
  6. Can the reservoir impoundment cause significant changes in fluid effort and pressure capable of affecting the seismotectonic environment?

Furthermore, three years before the lake is filled, at least one tri-axial geophone should be installed. At the beginning of the impoundment, at least three tri-axial geophones should be installed around the reservoir, to detect and locate microseisms (Barros et al., 2016).

Figure 3 presents a hypothetical scenario, in which ten tri-axial borehole geophones were installed, around the reservoir and on the dam body. Around the reservoir, the geophones can be installed at different depths, to prevent having a plane array, consequently increasing the sensitivity in depth and decreasing the epicenter location errors. The proposed depth will vary depending on the water table depth and relief restrictions. On the dam body, the borehole geophones are shallow, installed about 50 centimeters below surface (quasi non-invasive installation). All geophones can be integrated at the same seismic server (microseismic system), raising the system sensitivity. However, if it is of the owners’ interest, it is also possible to have two separated systems, considering the reservoir and the dam body as independent features.

Conventional (Passive) Microseismic

Considering an occurrence of a seismic event below the reservoir, the microseismic system will identify the P-wave and S-wave arrivals on the registers from all geophones. The seismograms will be associated with the event and will integrate the system database.

Picking the P-wave and S-wave arrivals, using a homogeneous velocity method, the system will estimate a distance radius from each sensor. The interception among the radius will indicate the event epicenter and, afterwards, the ground motion data will allow the estimation of other source parameters, such as seismic energy, moment and potency.

The conventional approach allows the processing and manipulation of this data, providing tools for a better understanding of the structure’s behavior after the occurrence of a seismic event. One of the most used tools is the generation of shake maps, in which the vibration data measured by each sensor is interpolated by the inverse distance weight method (IDW).

The characterization of an event, as an induced, will be possible using a statistic analysis of the historical data (Figure 5). The conventional approach builds a robust database, in which it is possible to characterize the seismic sources and observe whether this natural mechanism event was the first identified (1st question) or if the rate of occurrence of events has been increasing (3rd question).

After the statistical analysis and a positive answer to the questions presented by Barros et al. (2016), it is possible to characterize the seismic event as induced. It is important to highlight that an initial seismicity could be a result of instantaneous effects, such as the loading provided by the water column. The main event usually occurs after the total filling of the reservoir. This time frame between the start of the filling and the main event can take as long as three years (BARROS et al., 2016).

Seismic Ambient Noise Interferometry

The interferometric approach applied to dams’ monitoring is an innovative methodology, having been applied in an operational scale for the first time by Tetra Tech South America and the Institute of Mine Seismology in Brazil on tailings dams.

The seismic ambient noise interferometry uses an omnipresent source, known as ambient noise, to monitor wave propagation velocity changes in the medium. Using pairs of geophones, this method extracts the Green’s function between them. Green’s function is the wavefield recorded over time at one sensor when an impulsive source is triggered at the other sensor (PLANÈS et al., 2016). According to Wapenaar et al. (2009), interferometry refers to the principle of generating a new response by cross-correlating seismic observations on a receiver, being a sensor a virtual source.

The cross-correlation is computed periodically for all sensor’s pairs. From Tetra Tech’s experience, 10 minutes is frequent enough to indicate changes and sparse enough not to overwhelm servers and storage systems. From the first cross-correlation (cross reference correlation) the others are computed. Variations in transit-time between wave arrivals are evaluated. However, due to the greater sensitivity and robust signal, this verification is made in the portion of the seismogram called coda wave. The changes that are measured in the coda wave mostly represent changes in shear wave (S-wave) velocity (DE WIT; OLIVIER, 2018).

Shear waves are body waves associated with the shear stimulus of the propagation medium.

Thus, the interferometric approach monitors the ratio of stiffness modulus and density. The results provide a velocity change time series. The series trend is the result of the influence of several random components, which impact the values of stiffness and specific weight. Among them, pore pressure stands out for tailings and earth-filled dams, whilst fracturing stands out for concrete dams.

Considering the previously described scenario, with a monitoring system around the reservoir and on the dam body, the reservoir filling changes stress state, through the increase of both vertical stress and pore pressure.

From the hypothetical array indicated in Figure 3, taking into consideration the sensors pair 1 – 4 and having the change in pore pressure as the only variable, the hypothetical result would be given by the graph.


Microseismic monitoring technique can be divided into two approachs – passive microseismic and seismic ambient noise interferometry. Although not all suppliers are able to provide interferometry, its relation to stiffness modulus and specific weight provides valuable information for geotechnical engineers to manage their structures.

In addition, passive microseismic provides not only valuable information for the dam body, especially for mining dams, allowing owners to differentiate blasts from natural events. But also, passive microseismic enables owners to monitor seismicity in the reservoir and its surroundings, which is important to track the historical records and indicate possible induced seismicity, locally and regionally monitoring the geology and hydrogeology of the rock mass.

In conclusion, the microseismic monitoring system should ideally be conceived prior to the filling of the reservoir, to provide enough data to build a history regarding local seismicity, and should also consider the dam body, as it can be employed as a monitoring technique for its integrity.


  • Barros, L. et al. Sismicidade Induzida por Reservatório (SIR): casos brasileiros. Boletim da Sociedade Brasileira de Geofísica, v. 96, p. 16–21, 2016. In Portuguese.
  • Bohnhoff, M. et al. Passive seismic monitoring of natural and induced earthquakes: Case studies, future directions and socio-economic relevance. New Frontiers in Integrated Solid Earth Sciences, International Year of Planet Earth, 2010.
  • De Wit, T.; Olivier, G. Imaging and monitoring tailings dam walls with ambient seismic noise. Australian Centre for Geomechanics., p. 10, 2018.
  • Institute of Mine Seismology. Sensors, 2021. Available at: <>. Seen on: Jan. 3rd, 2021
  • Mendecki, A. J. A guide to routine seismic monitoring in mines. In: A handbook on rock engineering practice for tabular hard rock mines. Cape Town, South Africa: Safety in Mines Research Advisory Committee to the Department of Minerals and Energy of South Africa, 1999.
  • Planes, T. et al. Time-lapse monitoring of internal erosion in earthen dams and levees using ambient seismic noise. Géotechnique, v. 66, n. 4, p. 301–312, abr. 2016.
  • Rodrigues, C. T.; Oliviera, L. A. Microseismic technology for tailings dams monitoring. Water Power & Dam Construction, 29 set. 2020. Available at: <>
  • Wapenaar, K. et al. Tutorial on seismic interferometry: Part I – Basic principles and applications. Geophysics, 2009.