Application of 3D empirical travel times to routine event location

The unmodeled effects of lateral heterogeneity are a primary cause of event mislocation in routine methods. Approaches that account for the effect of lateral heterogeneity on travel times are commonly used in specialized studies, but some methods have recently been developed that are suitable for routine location. A variety of changes to the routine procedures have been suggested by other authors. These proposed changes are shown to be compatible to with the use of 3D travel time predictions. 3D empirical travel times are discussed in a routine context. This one step approach, based on direct use of a database of the most reliable previous observations, is shown to be suitable as an additional final step in routine procedures to give a supplementary high accuracy location. The empirical travel time approach is tested on a reference dataset of 155 ground truth events and shown to reduce epicentral mislocation by 45%. The approach is found to be particularly effective at reducing large mislocations. Improvement is demonstrated for every region represented in the dataset but particularly large improvements are shown for events in Europe, USA and Japan.