An aggregated template methodology: Novel automatic phase‐onset identification by template matching
ABSTRACT The precision of P‐ and S‐wave phase picking strongly determines the precision of earthquake locations, but such picking can be challenging in the case of emergent signals, large data sets or temporally varying seismic networks. To overcome these challenges, we have developed the concept of...
| Published in: | Geophysical Prospecting |
|---|---|
| Main Authors: | , , |
| Format: | Article in Journal/Newspaper |
| Language: | English |
| Published: |
Wiley
2021
|
| Subjects: | |
| Online Access: | http://dx.doi.org/10.1111/1365-2478.13103 https://onlinelibrary.wiley.com/doi/pdf/10.1111/1365-2478.13103 https://onlinelibrary.wiley.com/doi/full-xml/10.1111/1365-2478.13103 |
| Summary: | ABSTRACT The precision of P‐ and S‐wave phase picking strongly determines the precision of earthquake locations, but such picking can be challenging in the case of emergent signals, large data sets or temporally varying seismic networks. To overcome these challenges, we have developed the concept of an aggregated template to perform automatic picking of the P‐ and S‐wave phases. An aggregated template is defined as a representative event for a small area, built by aggregating the best signal‐to‐noise‐ratio seismic traces from events with similar waveforms (i.e. multiplet events). A template matching procedure, based on the cross‐correlation between an aggregated template and an unpicked event, automatically determines the unpicked event P‐ and S‐wave phases. This method enables (1) consistent and accurate P‐ and S‐wave phase picking and (2) reduces processing time relative to traditional template matching by using a clustering method that finds the most representative templates for a region, and thus limiting the required number of templates. We established two parameters to weight the picking precision: (1) the cross‐correlation between the aggregated template and the unpicked event and (2) the number of P‐ and S‐wave picks determined per event. We tested this method on 2100 events recorded in the south‐west of Iceland. Nineteen aggregated templates have been defined and used to automatically pick ∼65% of the complete event catalogue with an accuracy within the range of the manual picking uncertainty. These automatically picked events can then be used for event location, even when characterized by low magnitude, low signal to noise ratios and with emergent P‐wave signals. |
|---|