Example Cluster: Jordan & Sverdrup Region A

In Jordan and Sverdrup (1981) the hypocentroidal decomposition method is applied to three small clusters of earthquakes in the south-central Pacific (Line Islands), labeled regions A, B and C. This example cluster is their Region A.

If you are not yet familiar with the naming conventions for clusters in mloc, please review the guidelines. The cluster name is “jsa” and the series number is “5”. The output of only one run of mloc is included, so the basename is jsa5.1.

This is a very simple cluster with only four events. It cannot be calibrated because there is no close-in arrival time data or independent knowledge of the locations, so it is a basic teleseismic relocation that only provides an improved estimate of the relative locations.

The results of running mloc on the jsa5 cluster, either the output files included in the distribution or from a run you do yourself, will vary somewhat from what is presented in Jordan and Sverdrup (1981) (J&S). This is due to differences in the arrival time dataset, the travel-time model and the weighting scheme:

• J&S repicked some arrival times themselves.
• J&S used the Herrin (1968) travel-time model (P only).
• J&S used first-arriving P arrivals only, even for cluster vectors. No Pn or PKP phases were used.
• Readings with residuals greater than 5 s were discarded by J&S.
• J&S used reading errors of 0.7 s and 1.0 s, based on subjective judgement of quality.

The dataset in our example consists only of the readings available from the ISC Bulletin. In the example jsa5.1 run the ak135 model is used to estimate the hypocentroid, using P arrivals between 30° and 90°. Default reading errors (0.5 s for P) are used. Cluster vectors are estimated with all available phases at all distances. The windowing algorithm (command wind) is based on the default values for spread of different phases. For the P phase, readings with residuals greater than 1.5 s will be rolled off and readings with residuals over 2 seconds will be dropped from the relocation. In subsequent runs of mloc, we would begin using empirical readings errors from the ~.rderr file of a previous run (command rfil) and observed values for the spread of different phases from the ~.ttsprd file of a previous run (command tfil).