Modelling limits to earthquake detection and source parameters assessment

Clarifying the theoretical limits on earthquake detection and reliable analysis of earthquake source parameters can help in optimizing the design of seismic networks for monitoring natural seismicity and induced seismicity related to georesources exploitation. It can also help in assessing whether reports on event detection and derived source properties are consistent with expected propagation of information from sources to receivers.

fastloc GmbH uses the methodology developed in Kwiatek and Ben-Zion (2016) to a) estimate the performance of passive seismic monitoring networks in detecting seismic events, and b) estimate expected quality of source parameters derived from seismic data recorded with the planned network. The methodology is based on synthetic calculations of seismic radiation from kinematic earthquake sources, and incorporates effects of source-receiver path effects, ambient seismic noise, and properties  of recording systems.

Use cases

  • Characterizing limits to earthquake detection and reliable assessment of seismic source parameters before installation of seismic network
  • Assessing performance of passive monitoring networks for the purpose of traffic light system operations

Method overview

Figure 1. Modelling scheme used to assess the detection limits and limits to reliable assessment of source parameters. The source is described by seismic moment (magnitude), stress drop, rupture velocity, and focal mechanism. The path effects are described by the source-receiver distance  and intrinsic attenuation. The noise is superimposed and the recorded waveform is affected by the transfer function of the sensor and sampling rate .

We used the modelling scheme developed in Kwiatek and Ben-Zion (2016) which is schematically presented in Figure 1. We model synthetic seismograms using defined set of source and path characteristics, add to the generated seismograms random motion based on a defined noise model and influence of recording system, and finally calculate ground motions and signal-to-noise ratio of P- and S-waves. These method allows us to examine how different sets of seismic source, path, sensor, and noise characteristics affect the amplitude and frequency content of hypothetical seismic events occurring at the investigated site. This, in turn, allow to characterize and optimize the passive seismic network monitoring in the considered project.

Input and output data

Recorded seismograms can be interpreted as convolutions of source, path, and sensor effects superposed with the ongoing ambient seismic noise. To generate synthetic seismograms, basic information about passive seismic network monitoring setup and medium properties are required. This includes:

  • site and passive seismic network monitoring setup data include information on anticipated location of the seismicity and proposed design of seismic network. This includes e.g. location of injection or exploitation (source of seismicity), location of seismic stations, type of seismic sensors (broadband, short-period etc.), specification of the acquisition system (e.g. sampling rate, dynamic range etc.).
  • information on geological medium in which the simulated waves will propagate include information on seismic velocities, rock density, attenuation and site effects.
  • information on noise could be added if actual noise spectra are available from already working seismic systems operating in the vicinity of the project site. Alternatively, the noise will be generated from earth noise spectra taking into account the situation at the site.
  • information regarding properties of seismic sources can be considered in synthetic waveform modelling if it is available. This includes e.g. expected kinematics of earthquakes and corresponding radiation pattern of waves if one knows that the seismicity will occur on fault of known orientation, static stress drop, rupture velocity

Based on the above information, the synthetic seismograms are generated for events with different magnitudes and optionally other source properties, and the signal-to-noise ratio as well as expected frequency content is estimated on stations forming the monitoring network. This results in information on earthquake detection limits, typically presented as a function of earthquake moment magnitude. as presented in Figure 2.

Figure 2. Sample detection thresholds from project F. Expected signal-to-noise (S/N) ratio are presented as a function of the moment magnitude. The transparent blue graph corresponds to S/N ratios for assumed P-wave source-receiver distances between 3 – 10 km (shorter-farther source-receiver distance). The corresponding S/N ratio for S-waves is shown using magenta. The filled circles mark observed dominant frequency of the signal.

From sample graph above one can see, that for the particular combination of path, sensor and noise that are site characteristics, the detection limits vary with magnitude and distance. In the presented case, events with magnitude -1 are detectable at 3km hypocentral distance. At 10 km distance, the detection threshold is increased to magnitude 0. As one can see, the modelling is performed separately for P and S waves. This is convenient for modelling the hypocenter location uncertainties, as one knows which phases are detectable at which distances. Results of simulation allows us to quantify whether events of particular characteristics will be detectable with the proposed passive seismic monitoring network and allows us to optimize the network further towards better location estimation quality.

Together with earthquake detection plot we typically estimate the frequency content of incoming waves. This allows to provide additional suggestions to the design of monitoring system: e.g. which type of sensors to use, and how to optimize the acquisition system to get the best quality seismic data (Figure 3) in terms of frequency band.

Figure 3. Relation between observed dominant frequency content calculated for P- and S-waves and S/N ratio. The transparent blue graph corresponds to S/N ratios of P-waves and source-receiver distances ranging 3-10 km. The corresponding S/N ratios for S waves and similar source-receiver distances is shown in magenta. The filled circles mark the moment magnitude.

This service is provided to the customer in a form of the report.

Contact / questions

For further inquiries regarding this service please contact .

Futher reading

Kwiatek, G. and Y. Ben-Zion (2016). Theoretical limits on detection and analysis of small earthquakes. Journal of Geophysical Research-Solid Earth 121, DOI: 10.1002/2016JB012908.