The Effect of Secondary Arrivals on Regional Earthquake Locations

Synthetic Tests, Part 2

Relocations of synthetic EDR and LANL data with gaussian noise added to all arrival times. Noise is also dependent on both arrival and model errors. Run 10 of 10 is shown. True synthetic origins as plotted as black dots. Velocity models IASP91 and SSB (Introduction) were used with and without defining S phases. Average distance from the true synthetic origins is shown in the graph in the lower right corner of each plot. Each column represents a different arrival error scheme. Column A have arrival errors as given in the EDR catalog and 2.0/3.0 seconds for LANL P/S picks, respectively. Column B have all P/S phases with errors of 2.0/3.0 seconds. Column C have equal errors for P and S phases (2.0 sec). Column D have "regional" errors: LANL picks - 1.0/2.0 sec for P/S; EDR picks - 2.0/3.0 sec for P/S. This "regional" weighting produces the most accurate and precise relocations, with the most consistent average distance for the 10 location runs. This is expected given that the gaussian noise in the arrival data is dependent on starting arrival error. The use of S phases in the locations also produces better location results.





EDR Catalog Errors
LANL Arrival Errors

2.0/3.0 s for P/S
All Arrival Errors
2.0/3.0 for P/S
All Arrival Errors Equal
2.0 s
EDR Arrival Errors
2.0/3.0 s for P/S
LANL Arrival Errors
1.0/2.0 s for P/S

Introduction | Catalog Locations with Secondary Phases | Ground Truth Constraints on Relocations Near a Large-Magnitude Event
Synthetic Tests Pt. 1 | Synthetic Tests Pt. 2 | Relocations of Actual Data | Relocations Using S Phases for the Asia Study Area | Summary