I.e consider an earthquake seismic record (seismograph) which obtained from a mid-level noisy station. In this case we have a significance picking error for both P and S wave. is there any relation (mathematical equation) between the seismograph noise and phase picking error?
It's very difficult to say, as, among others, two similar events recorded at the same station, at the same distance, but with different focal mechanisms will have a different P-wave onset, and could range from faintly emergent to very impulsive. Of course, intuitively, you would expect that a nosier station will lead to noisier measures. Remember that the energy of the P or S wave might be in a different frequency band than the noise, and properly filtering your signal could significantly reduce the picking error.
It is an intriguing question but at moment, in my opinion ,is without a definitive answer. Many attempts are in literature.
Some mathematical formalization and methods has been developed in exploration seismic: i.e. automatic pickers for P-waves.
Also in seismology several automatic pickings methods are developed and some picking error function are proposed but in my opinion they depend on from the data set and they not are generalizable.
In general more than the picking error we can define - a posteriori- the failure degree in picking operation.
Agreeing with Thomas, to pick a phase and/or to define a picking error function we should know the source and propagation processes (from which the waveform depends on and which are different for each source-station couple) but it is just what that we want to find by performing the picking and solving the inverse problem.
About this for you could be interesting the results of experiment about the "picking operation" performed by the project "Seismogram Picking Error from Analyst Review (SPEAR)".
http://www.bssaonline.org/content/99/5/2759.abstract
- Badges
- Science topic