|
Traditionally, assessments of seismic network detection capabilities are based upon assuming statistical models for the noise and signal distributions. These models include station corrections for signal attenuation and a combinational procedure to determine the detection threshold as a function of the number of phase detections required for reliable location.
In general, it is implicitly understood that any network will have a detection threshold that varies with time. It is important to retain such information along with the information on the average capability. However, the methods used in practical operation today make no attempt at specifying the time-dependency of the calculated threshold. For example, the noise models used in these capability assessments are not able to accommodate the effect of interfering signals, such as the coda of large earthquakes, which may cause the estimated thresholds to be significantly degraded at times. Furthermore, only a statistical capability assessment is achieved, with no time-dependent evaluation of when the possibility of undetected seismic events is particularly high, for example during unusual background noise conditions or outages of key stations.
The continuous threshold monitoring technique, which has been developed at NORSAR, is intended to address these issues. Basically, the difference between the threshold monitoring approach and traditional detection threshold estimation can be described as follows (assuming a statistical model with a given confidence level):
- The detection threshold is an estimate of the smallest hypothetical seismic event at a given site or in a given region that could possibly be detected (e.g. by 3 stations)
- Threshold monitoring provides an estimate of the largest hypothetical seismic event at a given site or in a given region that could possibly have occurred.
The two approaches are therefore complementary, and each provides useful information in the context of seismic monitoring. The threshold monitoring approach could be especially useful to identify time intervals when the possibility of significant “hidden” seismic events is particularly high, thus enabling the analyst to concentrate on such time intervals for extensive analysis. Furthermore, the method provides an upper limit of the magnitude of non-detected events, which could be useful e.g. to assess the maximum MS value for events for which no surface waves are detected.
In the articles below, we give an example of application of the threshold monitoring method to estimate the detection capability of available seismic stations to monitor the North Korean nuclear test site.
|
 |
|
|
|
